Optimizing the scheduling of farm machinery is essential to meet farmers’ requirements, minimize scheduling costs, and save time. This study focuses on scheduling farm machinery in multiple cooperatives across various regions, aiming to minimize scheduling costs and reduce scheduling time. Initially, a multi-constraint hybrid clustering algorithm is employed to assign farmland to each farm machinery cooperative by clustering before scheduling. Subsequently, an enhanced version of the Non-dominated Sorting Genetic Algorithm-Ⅱ (NSGA-Ⅱ) is proposed, integrating a local search strategy based on congestion-based neighborhood search and the Simulated Annealing (SA) algorithm to develop the NSGA-Ⅱ-SA algorithm. This hybrid multi-objective evolutionary algorithm effectively optimizes scheduling costs and time. The model’s validity and the algorithm’s superiority are demonstrated through a Web-based multi-region agricultural machine scheduling system and an example study. Experimental results show that the NSGA-Ⅱ-SA algorithm significantly reduces scheduling costs and time, as well as the number of scheduled farm machines, outperforming other algorithms with reductions of 9.8%, 3.1%, and 8.7% in total scheduling costs, and 12.5%, 13.4%, and 11.6% in total scheduling time. This research establishes a theoretical framework for multi-region agricultural machine scheduling in hilly and mountainous areas, enhancing agricultural production efficiency.

To address the issues of low response efficiency, poor timeliness, and high scheduling costs in the collaborative operation of harvesting and transportation, a cooperative scheduling model was constructed in this paper with the objectives of minimizing both scheduling costs and time. The model aimed to solve the challenge of coordinating harvesters and grain transport vehicles of different capacities under time window constraints, thereby completing the harvesting and transportation tasks across multiple fields efficiently. First, to improve the response efficiency and timeliness of harvesters and grain transport vehicles in complex operating environments, a task allocation method based on priority strategy and a response scheme were designed to enhance the efficiency of cooperative operations. Secondly, an Improved Marine Predators Algorithm (IMPA) was proposed by incorporating a radial transformation matrix, a dynamic search mechanism, and an enhanced opposition-based learning strategy to strengthen the overall optimization capability of the scheduling scheme. Additionally, non-dominated sorting and crowding distance calculation were integrated into the algorithm to optimize the solution set, leading to a more optimal scheduling plan. Simulation results demonstrated that, compared to the Non-dominated Sorting Genetic Algorithms-II (NSGA-II) and the Multi-objective Harris Hawks Optimization (MOHHO) presented in the literature, the proposed model significantly reduced scheduling costs and time, with a reduction of 7% and 2.9% in scheduling costs and a reduction of 2.6% and 8.3% in scheduling time, respectively. The results indicated that IMPA not only reduced scheduling costs and time but also enhanced the response capability and operational efficiency of cooperative scheduling, offering significant practical value in solving the collaborative scheduling problem for multi-field harvesting and transportation.