Abstract: Sudden leaks in oil pipelines situated in mountainous areas present significant challenges for urgent repairs due to complex terrain, limited road access, and highly sensitive environmental conditions. To address these challenges, this paper presents an improved ant colony algorithm(I-ACO). This approach introduces a composite cost function that integrates terrain resistance, leakage risk, and environmental sensitivity. Additionally, the function incorporates quantified multidimensional constraints, dynamic pheromone evaporation rates, prior preferences for existing roads, and local search strategies. As a result, the algorithm is enhanced to improve the convergence rate while reducing the likelihood of converging to local optima. The simulation results indicated that the convergence rate of the I-ACO was 60.0% higher and the rate of traversing sensitive areas was 57.1% lower in 20×20 grid tests compared to traditional ant colony algorithms, across multiple scenarios involving existing roads and environmentally sensitive areas. In a 200×200 real-scenario terrain mapping, the convergence rate increased by 17.5%, and the likelihood of selecting paths through flat, low-risk areas rose, enhancing the algorithm's adaptability to urgent repair needs. These findings offer an efficient and reliable path optimization approach for urgent repairs in response to oil pipeline leaks in mountainous areas.