Abstract: Relying on the distributed manufacturing concept and the flexible deployment characteristics of Mobile Factories (MF), the mobile supply chain enables resource sharing and rapid response in scenarios with multiple production sites and multiple customers. However, the coupled optimization challenges of order assignment, production scheduling, and MF routing significantly increase the scheduling complexity, and existing methods struggle to efficiently balance solution quality and solution efficiency. Aiming at the Mobile Supply Chain Scheduling (MSCS) problem considering multi-production-site and multi-customer scenarios, this paper establishes a mathematical model with the optimization objectives of minimizing makespan and total cost, and proposes an improved NSGA-II algorithm for solving it. The algorithm incorporates four key improvements: First, a four-layer vector encoding structure based on orders, production sites, and mobile factories is designed to accommodate the four coupled decision-making processes of the problem. Second, an objective-oriented heuristic population initialization strategy is proposed to enhance population quality and accelerate algorithm convergence. Third, crossover operators targeting two types of vectors are designed to improve the algorithm's global exploration capability, and four critical path-based neighborhood search operators are developed to strengthen its local exploitation capability. In the experimental section, based on a practical case of a chemical enterprise, 30 test instances covering three scales (small, medium, and large) are constructed. Three groups of experiments, including parameter calibration, improvement verification, and algorithm comparison, are conducted, which verify the effectiveness and superiority of the proposed algorithm in solving the coordinated assignment and production routing problem of the mobile supply chain.