The objective of this study is to make a study on dealing with uncertainty in the design of river training works. Based on this study, a method and several instruments have been developed that aim to improve a full and consistent use of a probabilistic design approach. This study is divided into three parts. The outline is rather unconventional for a civil engineering thesis and some of the subjects treated are more related to Information Logistics and Data Management than to typical constructional aspects of civil engineering structures. But since information has become a vital asset for almost every company, in particular for an engineering consultant, it seems a very relevant aspect to study. The first section is a general introduction to the types of uncertainty that have to be dealt with in the design of engineering structures and to the method that is used to incorporate these uncertainties: the probabilistic design approach. In a brief discussion of this approach, two of the main constraints for this approach are introduced: 1. One of the main difficulties is the statistical description of parameters in case of limited available data. These descriptions are used in the probabilistic calculations. If hardly any information is available or only a 'practical' design model is available, does it make sense to apply a probabilistic design approach? 2. Another important aspect is that probabilistic calculations and sensitivity analyses can be very timeconsuming. Because many design studies have a limited time-schedule, this can prevent a full use of these methods. Is this inevitable or could certain methods and design instruments be developed that improve this applicabihty? In the second section part of the first question is analysed by comparing the design results of three similar projects, in which river training works were designed. Two of these projects used a probabilistic design approach, the third one a deterministic approach. In a deterministic approach the safety is not quantified, in a probabilistic design approach, the probabilities of failure of all components of the structure are explicitly determined. The design dimensions of the structures designed in the three projects were likely to be different because of differences in natural conditions, required safety and the design models used. The objectives of this section are to study how the probabibstic design method was applied in the two projects, to study how the uncertainty of the parameters were determined and to quantify the influence of the differences in natural conditions, required safety and used models on the differences in the final designs. In the third section, the design process of an ongoing project is analysed in order to study the second question posed in section I. The ongoing project is the Mekong River Bank Erosion Study which will be used as a reference and case study. The objective of this section is to develop a method and instruments to improve the incorporation of uncertainty in the design. First the methods of determining the uncertainties of observed and derived parameters are studied. Next, the way in which the information about these parameters is managed, communicated and used is discussed. Finally, a design model is proposed in which probabilistic calculations can be made on every required level of detail, using one single software program.