Reliability is becoming a more and more important performance characteristic of road networks and transport services. Travel time reliability was found to be one of the most important factors for route choices of travelers. Drivers often complain that it is the unpredictability of travel time that they dislike most of a journey in a congested city or area. Network design and some traffic control measure or ITS, such as Advanced Traveler Information System (ATIS) and Incident Management (IM) have been proposed to improve road network reliability. Road pricing has been advocated as a potentially powertlil travel demand management (TDM) strategy capable of significantly influencing travel demand characteristics through the network. Generally, the policy objecfives of road pricing could be summarized as follows: managing demand, optimizing congestion level, reducing environmental impacts, maximizing social welfare gains, raising revenues to recoup maintenance cost and construction cost, etc. To our best knowledge, no researchers have advocated improving network travel time reliability as an objective of road pricing. Some researchers suggested evaluating road pricing from a network reliability view. These suggestions give some incentives that road pricing could make contributions to network reliability. This study investigates road pricing as an instrument to improve network travel time reliability, which is induced by link capacity variations and short-term origin-destination demand fluctuations. The network reliability is specified as a function of standard deviations of route travel times. It is assumed in this study that the road authority aims to optimize network reliability by setting tolls in a network design problem. Travelers are influenced by these tolls and make route and trip decisions by considering travel times and tolls (not taking the trip reliability into account). Theoretical analyses using static models with elastic demand are performed to analyze the influence of road pricing on road network reliability. In the theoretical analyses, capacity variation is assumed as the major influencing factor on the network travel time unreliability. An important assumption is made in the theoretical analyses, being that travelers make their route choices based on the average travel cost. Travel time and link capacity are stochastic variables, while route flow is deterministic for each charge level. The theoretical analyses start with the analysis on a single route network. Then a tworoute network with one link charged is analyzed using deterministic user-equilibrium (DUE) and stochastic assignment. Finally theoretical analyses on a general network are performed and a formula to calculate route travel time unreliability is derived based on oiu- assumptions mentioned previously. This formula is utilized in an application on a hypothetical five-link three-route network to investigate the impacts of road pricing on network reliability. The theoretical analyses show that road pricing may improve network reliability and that network reliability depends heavily on charge levels. Theoretical analysis shows that decreased OD demand and route flow switches, caused by implementing tolls, influence network reliability. Depending on the combination of the effects of decreased OD demand and route shifts, network reliability in case of elastic demand can either be improved when the effect of OD demand decreases is stronger than the effects of route shifts, or becomes worse when the influence of the route shifts is stronger than the influence of decreased OD demand. Whereas the theoretical analyses are performed using static models and some assumptions had to be made, a dynamic simulation-based method using the dynamic model INDY for the five-link network is utilized to investigate the influence of road pricing on network reliability more appropriately. Using the dynamic simulation-based approach, capacity variations and short-term origin-destination demand fluctuations are considered as the most important factors affecting travel time variability. Fixed OD demand and elastic OD demand are taken into account respectively. Four cases are analyzed in the simulation-based analyses; the fixed demand case with capacity variations, the fixed demand case with demand fluctuations, the elastic demand case with capacity variations and the elastic demand case with demand fluctuations. With fixed demand, route shift is the major causal factor of the variability of network travel time reliability with respect to charge levels. The results from the simulafion-based method show that in the fixed demand cases with the increase of the charge levels, route flow switches become less and less. Network reliability can be improved with tolls. With elastic demand, both decreased OD demand and route shifts play an important role in network reliability. Decreased OD demand leads to positive effects on network reliability. Frequent route shifts is the essential cause of network travel time unreliability for the networks with tolling systems, and network reliability will be improved i f there are less route flow switches. Combining the effects of decreased OD demand and route shifts, network reliability may either be improved or becomes worse, depending on the charge levels. The optimal charge from road authority's view can be determined, aiming to optimize network reliability. With the optimal charge, people take longer but more reliable routes. It turns out that there is a trade-off between the average route travel time and the route travel time reliability. In a normal situation (without tolls), both capacity variations and demand fluctuations have great influence on travel time variability, since these two factors lead to stochastic route flow shifts. Demand fluctuations have greater influence on travel time variability than capacity variation does. Because demand fluctuations have direct influence on route flows, which leads to direct travel time variability. The objective of optimizing network travel time reliability proposed in this study can contribute to the charging system design, which is a bi-Ievel optimization problem. Road authority, as the leader, tries to optimize network reliability by setting tolls. Travelers, as the followers, try to minimize their own travel cost, which is the sum of travel time and tolls. The objective function can be used to optimize charge levels or search optimal links for charging for fixed charging systems and for variable charging systems. Only a small hypothetical network with a single OD pair has been analyzed in this study. For this small network, we conclude that road pricing can improve network travel time reliability. For more general networks, the objective of optimizing network reliability can be employed to determine the optimal charge level and to optimize the charging system design.