The passengers' flow and station dwell time estimation are important tasks for mass transit planning. However, classical methods are difficult to apply into some practical achievements. This paper presents a new approach that models passengers' flow and its effect on passenger alighting and boarding time in mass transportation systems in the presence of uncertainties. The applied technique combines origin destination matrices approach with the application of artificial intelligence. This new approach allows the inclusion of some intuitive knowledge provided by a fuzzy logic inference motor to predict the flow demand of passengers' trips, alighting and boarding time passenger cars in explicit stations.