The traditional Pelican Optimization Algorithm (POA) has problems such as low convergence accuracy and a tendency to get trapped in local optima. To address these issues, this paper proposes an improved Pelican Optimization Algorithm, namely IPOA. Firstly, this algorithm uses the opposition-based learning strategy to initialize the population, which improves the quality of the initial population. Secondly, it incorporates the idea of balancing individuals to balance the optimal individuals, and introduces the Levy flight strategy to increase the diversity of the search. This helps the algorithm to escape from local optimal solutions and explore the global search space more effectively. Finally, a nonlinear inertia factor is introduced, endowing the algorithm with stronger exploration ability and global optimization ability, which also contributes to getting out of local optima. Through comparisons with other intelligent optimization algorithms, the results show that the performance of the IPOA algorithm outperforms other intelligent optimization algorithms. When the IPOA algorithm is applied to the path planning of substation inspection robots, it has advantages over other algorithms, verifying that this algorithm has certain engineering capabilities.