Emergency response in urban gas pipeline networks is highly sensitive to stochastic traffic conditions, which introduce substantial uncertainty in crew travel times to leakage sites. Existing facility location models typically rely on predefined candidate sites and deterministic travel assumptions, limiting their ability to capture full-cycle dynamic recovery processes under random leakage events and traffic congestion. This study develops a candidate-free location optimization framework for repair station siting under stochastic road resistance conditions, aiming to characterize spatial variability in emergency response capability. The framework integrates a candidate-free facility location model with a hybrid greedy–Monte Carlo solution strategy to optimize station layouts across network-wide stochastic scenarios. Coverage reliability, response time, and construction cost are jointly considered to support robust siting decisions. A case study based on the real road and gas pipeline networks of City H demonstrates the effectiveness of the proposed approach. Across 20,000 stochastic road resistance scenarios, the optimized layout achieves an average service coverage rate of 97.77% within the specified response time threshold, while maintaining stable performance under variability. Although increasing the number of stations enhances response capability, the improvement exhibits clear diminishing marginal returns. These findings provide quantitative guidance for determining cost-effective station scale and prioritizing core hub locations under uncertainty. The proposed framework offers a structured decision-support tool for resilience-oriented planning, prioritization of critical segments, and evaluation of emergency response and maintenance strategies in urban gas pipeline systems.