Shared spaces have been catching interest as a solution for traffic calming for the last two decades. Hence, simulation models that accurately replicate cyclist behavior in mixed traffic are important. Cyclists, however received less interest in the available shared space models due to the complex dynamics of the bicycle. The available shared space models that either consider only pedestrians and cars as main road users or do not account for the bicycle kinematics effect on its movement behaviour. This paper presents an agent-based cyclist model for shared space traffic simulation. Specifically, the focus of this paper is a conceptual model that incorporates the rider cognition and the bike as two connected entities governing the cyclist agent motion. This concept is incorporated in our model to simulate cyclist free-flow motion and one-to-one interactions between cyclists and pedestrians. The novelty of our model lies 1) applying social forces to achieve autonomous navigation of the cyclist agent, and 2) preserving the kinematic characteristics of the bicycle during motion by applying a steering decision method. We utilize a publicly available cyclist-focused dataset to first calculate and understand the dynamic parameters of the bicycle, including speed, acceleration, and turning rate. We then validate the model’s performance through simulation experiments, as well as comparisons with ground-truth trajectories. The results so far reveal that applying steering behavior of the cyclist to the model can reproduce cyclist’s trajectories with comparable accuracy via simulation.