In order to improve the performance of sine cosine algorithm (SCA) with poor local search ability, an sine cosine optimization algorithm with self-learning strategy and Lévy flight (SCASL) was proposed. Firstly, the self-learning strategy and nonlinear weight factor of sine cosine algorithm was proposed, so that the search individual could remember its historical optimal position, which guided the individual to update its location in the optimization process, thus improving the local search ability of SCA. When the search was stagnant, the stagnation perturbation strategy based on Lévy flight was adopted to jump out of local optimum so as to improve the local optimum avoidance ability. Based on 13 classic benchmark functions, the numerical simulation was conducted and the results show that SCASL has higher computational efficiency, convergence accuracy and stronger local optimum avoidance ability compared with standard SCA and other state-of-the-art optimization algorithms such as SSA,VCS,WOA and GSA. The simulation results of unmanned combat aircraft flight path plan show that for the battlefield environment with six enemy threat sources, SCASL can stably obtain less expensive flight path than the standard SCA.Therefore, the proposed SCASL has better optimization performance.