Predicting traffic flow is one of the fundamental needs for comfortable travel, but this task is challenging in vehicular cyber-physical systems because of ever-increasing uncertain traffic big data. Although live data with outstanding performance recently have become popular, most existing models for traffic flow prediction are fully deterministic and shed no light on data uncertainty. Also, they are many inventories in automobile industries to design and build safety measures for automobiles, but traffic accidents are unavoidable. There is a huge number of accidents prevailing in all urban areas. In this study, a novel dynamic approach is proposed for predicting citywide traffic flow based on weather. The proposed system utilizes the neuro-wavelet algorithm to select the required features for traffic prediction. The system also proposes the K Nearest Neighbor algorithm to predict the accuracy of accidents occurred in urban regions.