Abstract: Order allocation and route scheduling are often required in agricultural machinery service systems in modern agriculture. Intelligent cooperation and large-scale operation can be accelerated in practical scheduling scenarios. Resource utilization efficiency and service quality can also be improved with operational cost savings. However, suboptimal solutions are attributed to the coupled order allocation decisions and route planning. It is often required to fully meet the requirements of large scale, highly complex environments in the agricultural services. In this study, a hierarchical multi-objective optimization was proposed to integrate scheduling of order allocation and vehicle routing in distributed agricultural machinery service considering Farmers' Satisfaction. Multi-depot order allocation and internal route planning were also combined into a decision-making framework. A bi-level optimization structure was constructed from a system-level perspective. Among them, the upper level was used to allocate the orders among multiple agricultural machinery depots, while the lower level was used to optimize the route planning and machinery scheduling within each individual depot. Thereby, the interaction between allocation decisions and routing performance was explicitly captured after optimization. A multi-objective order allocation model was also formulated at the upper level. The total scheduling cost was simultaneously minimized to meet the farmer's satisfaction under operational constraints, including machinery capacity limitations, service time windows, and task continuity. A NSGA-II multi-objective evolutionary algorithm was developed to solve the upper-level model. Problem-oriented genetic operators and solution evaluation were incorporated to enhance convergence performance, solution diversity, and overall stability. A route planning algorithm was designed to further optimize the internal scheduling at the lower level after the upper-level allocation. A dual-layer chromosome encoding was adopted to integrate heuristic initialization, adaptive evolutionary operators, and local search mechanisms. Route feasibility and convergence efficiency were effectively improved under the more complex task. Simulation experiments were conducted to evaluate the effectiveness and practicality of the hierarchical optimization approach. Real scheduling demands were collected from an agricultural machinery cooperative in Henan Province, China. Multiple order scales and complex services were selected to compare conventional fixed allocation and classical algorithms as benchmarks. The experimental results demonstrate that the collaborative optimization consistently outperformed the fixed allocation under different order scales. Average scheduling cost was reduced from 2.88% to 5.25%, with a high level of service satisfaction. Farmer satisfaction increased by up to 18.86% in medium-scale order scenarios, indicating the effective balance of economic efficiency and service quality. Furthermore, the hypervolume metrics showed that the improved NSGA-II algorithm exhibited superior performance in both convergence and diversity, compared with the rest of the multi-objective evolutionary algorithms, indicating stronger global search and solution set robustness. In addition, the internal route planning was quantitatively analyzed for consistent cost optimization over all order scales. The maximum cost was reduced by 8.32% in large-scale task scenarios. The better performance of the route planning algorithm was achieved as the problem scale and complexity increased. The high adaptability and scalability outperformed conventional genetic algorithms in terms of cost efficiency and solution stability. Moreover, the average satisfaction level was improved in most order-scale scenarios, with the maximum improvement exceeding 40%. The local search mechanisms and adaptive evolution also reduced the service delays for the scheduling timeliness. Overall, the hierarchical multi-objective optimization framework effectively captured the coupling between order allocation and route planning. Coordinated optimization was realized at various decision levels suitable for the intelligent scheduling of multi-depot and multi-type agricultural machinery. The findings can provide strong support for the intelligent scheduling of agricultural machinery.