The performance of the NS and ProRail is measured by, among others, passenger punctuality. With the introduction of the OV-chipkaart (the smart card of the Dutch public transport) new data of passengers became available. Therefore the NS developed a new method of calculating passenger punctuality. Based on the check-in and checkout locations and the check-in time, for every passenger a promised journey (in Dutch: Reisbelofte) can be determined. This promised journey is the travel options with the earliest arrival time. Based on the realisation data of the trains and the checkout time, the delay of every passenger can be calculated. In this method of calculation more than 20 assumptions are made. In this research all assumptions are analysed. If possible the impact of each assumption on the passenger punctuality is calculated. The result is a possible error of approximately 2 percentage points. To improve the passenger punctuality two models describing passenger punctuality were analysed as well. The first model is developed at ProRail and the NS, the second model is proposed in this research. Both models try to describe passenger punctuality based on process indicators: performance indicators of the business unit Transport control and Traffic control. For both models the correlations between the various indicators are calculated. The results are comparable and for both models the strongest correlations are only moderate at best. The process indicators are focussed on delayed or cancelled trains while passenger punctuality describes delays of passengers. In a case study of a smaller area of the Dutch rail network, every correlation is calculated again. Some results differ significantly from the correlations in the entire rail network. This implies that localised steering to improve passenger punctuality might be more effective.