We introduce a probabilistic modeling approach for pedestrian speed density relationship. It is motivated by a high scatter in real data that precludes the use of traditional equilibrium relationships. To characterize the observed pattern, we relax the homogeneity assumption of equilibrium relations and propose a multiclass model. In addition to the general modeling framework, we also present some concrete model specifications. Real data are utilized to test the performance of the approach. The approach is able to reveal fundamental properties causing the heterogeneity in population and describe their impact on pedestrian movement. We also show the advantages of the proposed approach compared with approaches from the literature. The proposed model is flexible, and it provides richer information than traditional models.