In order to solve the problem of insufficient global search ability of original bat algorithm, an improved bat algorithm (IBA-FCS) was proposed, which effectively improved the global search ability of the algorithm by designing pulse frequency conversion strategy, adaptive local search strategy and mutation mechanism. The optimization results based on classical test functions showed that compared with particle swarm optimization algorithm, original bat algorithm and other improved bat algorithms, IBA-FCS algorithm had better optimization performance. Aiming at the track planning problem of agricultural UAV, combined with the three-dimensional terrain data of mountain orchard flight environment, the safety track planning model of agricultural UAV was constructed, and the flight cost function with multifactor constraints was designed;at the same time, the solution space of the track planning model was transformed from Cartesian coordinate system to cylindrical coordinate system to further improve the optimization efficiency of IBA-FCS algorithm, so as to obtain a better track planning scheme. Simulation results showed that in multiple missions with different numbers of obstacles, the fitness of IBA-FCS algorithm was 20.3355% lower than that of original BA algorithm, and the fitness of IBA-FCS algorithm based on cylindrical coordinate space was 4.6127% lower than that of IBA-FCS algorithm based on Cartesian coordinate space. The results of field experiments showed that the track planning scheme based on IBA-FCS algorithm can converge to the shortest track, which further verified the feasibility and effectiveness of applying improved bat algorithm and cylindrical coordinate system to agricultural UAV safety track planning in mountain orchard static obstacle environment.