To accommodate the continuous growth of electricity demand, the number of new and expanded substations being built by State Grid companies across provinces and cities is steadily increasing. As a result, some local power supply companies are facing staff shortages and heavy daily inspection workloads. To reduce the burden and increase the efficiency for frontline inspection and maintenance staff, this paper proposes a two-stage planning method of substation inspection path based on improved gold rush optimizer. First, an enhancement strategy is introduced for the emerging gold rush optimizer. The search efficiency and global optimization ability are improved by incorporating the free exploration behavior of gold miners and a cosine-based convergence factor, making it suitable for solving the subsequent two-stage substation inspection path planning model. Second, two substation inspection models are constructed: “multi-team work together” and “single-team work”, and the inspection parameters such as maximum route duration, type of inspection, and number of vehicles are preset. The improved gold rush optimizer is used to optimize and identify the benchmark work model. Next, based on the benchmark work model, the distinct advantages of the two work modes are leveraged to perform secondary optimization on each route, ultimately determining the substation inspection path that minimizes total time across the inspection cycle. Finally, simulation analyses are conducted using real-world substation work zones and centralized control stations. The results verify the effectiveness and general applicability of the proposed planning method.