Aimed at the problems that search ability is inadequate in search, convergence is slow at speed, and susceptible to local optima in the intelligent optimization algorithm for solving the UAV 3D flight planning problem, an Osprey Strategy Snake Optimizer (OSSO) is proposed. Firstly, Bernoulli chaotic mapping is introduced to initialize the population, expand the individual search range, and enrich the diversity of the population; Secondly, the search strategy is improved in combination with the ideas of submerged predations, stochastic step and precise mining in the Osprey Strategy Snake Optimizer, and the global search capability is enhanced; And then, the dynamic opposition-based learning is utilized for updating population, balancing the algorithm’s global exploration and local mined ability, and improving the algorithm’s ability to deal with local optima. Finally, two 3D models are constructed by the function method and elevation data respectively, and the simulation experiment is performed by taking the length of trajectory, the distance in threat zone and the drone physical constraints as the judging indexes. The experimental results show that the OSSO algorithm is rugged, and good in stability and in effectiveness in solving the three-dimensional track planning problems.