Abstract: To solve the problem of intelligent connected vehicles being prone to traffic accidents due to environmental weather conditions, a cellular automata traffic flow model integrating multi-source information was proposed. The neural networks were utilized to fuse multi-source information flows and high-precision map data in cloud environments and to achieve dynamic map updates and provide more accurate path planning for autonomous vehicles for effectively avoiding traffic accidents. Based on the multi-source information model, a two-lane traffic flow cellular automata model was constructed and compared with the manual driving model to demonstrate the superiority of the proposed model in terms of safety distance model, following rules and lane changing rules. The simulation was completed by MATLAB. The results show that the proposed model can increase traffic flow and reduce congestion time with achieving stable state in relatively short period of time, and the road traffic rate can be improved by 27%.