Unmanned aerial vehicle (UAV) path planning plays an important role in power systems. In order to address the challenge in UAV path planning, an improved crested porcupine optimizer (ICPO) combining the Cauchy inverse cumulative distribution function and JAYA algorithm is proposed in this paper. First, the traditional random initialization is replaced by sine chaotic mapping, making the initial population more evenly distributed in the search space and improving the quality of the initial solution. Since the global search ability of the crested porcupine optimizer (CPO) is limited, the Cauchy inverse cumulative distribution strategy is introduced. In addition, as CPO is prone to fall into local optima in later stages, a weighted JAYA-CPO attack strategy is proposed to balance the global exploration and local exploitation, thereby improving the algorithm's ability to escape from local optima. Finally, ICPO is compared with another 10 algorithms on the cec2017 and cec2020 test sets. The experimental results show that ICPO has excellent competitiveness and optimization performance. The ICPO algorithm is applied to the path planning problem of power inspection UAV and is compared with four algorithms. The results show that the algorithm can generate more feasible path trajectories across two terrains with varying complexity, demonstrating the effectiveness and significance of the ICPO algorithm for UAV power inspection path planning.