Abstract: To address challenges in multiple unmanned aerial vehicle（UAVs）agricultural mapping such as poor path coordination, weak obstacle avoidance capability in complex farmland environments, and difficulty in balancing mapping accuracy and efficiency, a dual approach was proposed through mathematical modeling and intelligent algorithm optimization．A large-scale farmland was selected as research scenario, by integrating crop type zoning, field heterogeneous obstacles, and differentiated mapping requirements to construct a multi-constraint and multi-objective cooperative path planning model．Building upon conventional genetic algorithms, an improved genetic algorithm was developed by incorporating a crop zoning weight matrix to optimize fitness function and designing a region-continuity crossover operator．Simulations were conducted in a 10 km×10 km agricultural planting area, comparing improved genetic algorithm with standard genetic algorithm（SGA）and non-dominated sorting genetic algorithm II（NSGA-II）．Results demonstrated that improved genetic algorithm significantly achieved an average total path length of 31.8±1.2 km, with coverage completeness reaching 99.3%±0.2%, and obstacle avoidance success rate at 100% while maintaining convergence speed at an average of 105 generations．Results significantly outperformed comparison algorithms, achieving a 100% resolution compliance rate in facility agriculture zones．An efficient and reliable technical framework for multi-UAVs cooperative mapping in precision agriculture scenarios was provided．