Abstract: Affected by the traffic volume during peak periods, large-scale queuing of vehicles is prone to occur during the transportation and scheduling of fresh agricultural products to be processed. In order to solve the vehicle scheduling problem for transporting fresh agricultural products to be processed under the influence of time-varying traffic, and to discuss the impact of vehicle speed changes on vehicle queuing and unloading, a time-varying speed characterization model with a multi-peak inverse Gaussian distribution is constructed to describe the impact of traffic conditions on vehicle speed. A time-varying speed optimization model for fresh agricultural products to be processed considering the impact of traffic is proposed. An improved genetic algorithm that combines a forward continuous crossover operator and an adaptive differential mutation operator is designed. Integrating the Gaussian distribution queuing time backward mechanism and using the binary search algorithm to accurately search the departure time. The results show that the time-varying speed model with multi-peak inverse Gaussian distribution can better describe the time-varying characteristics of simulated transportation vehicle speeds. The proposed mathematical optimization model and algorithm can effectively generate vehicle scheduling plans, which is of practical significance to ensuring the quality of raw material supply and improving logistics efficiency.