The urban public transport system in Casablanca, one of the major Moroccan cities, offers different alternatives. But the system is not efficient enough. Transportation options are rather frustrating, time-consuming, and sometimes risky. Many residents experience long and expensive commutes in often uncomfortable and unsafe vehicles through heavy traffic. In such a context, simulation can help to support urban mobility decision-makers in the task of shaping transportation initiatives. In this paper, we focus on improving the urban mobility system using an agent-based model. It offers the ability to capture and reproduce the mobility behavior of commuters. We create an initial scenario that represents the actual urban situation in Casablanca city. It is developed by creating synthetic populations and gathering other necessary data before integrating them into MATSim and running the simulation. This scenario will serve as a baseline to implement and design new models in our area of study.