The basic lion swarm algorithm is associated with low search efficiency and insufficient diversity. Thus, in this study, we propose a Sin chaotic population initialization operation to improve the quality of the initial solution of the algorithm. We also introduce an adjustment factor to improve the diversity of the algorithm. The directional constraint function increases the search accuracy and convergence rate of the algorithm, resolving the issue of path planning. We also propose the lion structure of two populations and improve the search ability of the algorithm through the mutual cooperation of differentiated populations. Path smoothing is achieved using the fourth-order Bessel curve. Finally, the improved lion swarm algorithm is demonstrated using test simulation. The performance of the proposed algorithm is significantly improved compared with basic lion swarm optimization, gray wolf optimization, particle swarm optimization, and genetic algorithms. Our findings show that the path planned by the improved lion swarm optimization is reduced by 5.67% on average, and the running time is reduced by 8.82% compared with the other studied algorithms.