According to the reality problem of city logistics distribution under the condition of traffic restrictions, we propose the multi-stage cross-regional joint distribution strategy. Considering the time constraints, we divide the customer service area by the customer time window and customer location, and launch distribution services in period of time and stages. Fully considering the distribution distance, vehicle-mounted rate, speed and other factors, and taking the fuel as the variable cost and vehicle launch costs as fixed cost, we design the optimization goal of minimizing the sum of total cost, and built a multi-stage cross-regional joint distribution model under the traffic constrictions. According to the characteristics of the model, an adaptive genetic algorithm is proposed. Finally, the example is given to illustrate the feasibility effectiveness of the proposed algorithm.