Abstract: Objective: To investigate the association between ambient temperature and the lethality of suicide methods, and to evaluate the potential effect modification by age, sex and season. Methods: A retrospective analysis was conducted on suicide-related emergency records from the 120 dispatch system of Wuhan Emergency Medical Center between June 2023 and May 2025, which were matched with concurrent meteorological data. The suicide methods were classified as high-lethality and low-lethality based on case fatality rate. Binary logistic regression was used to assess the association between daily mean ambient temperature and the high-lethality suicide methods, with adjustment for age and sex. Subgroup analyses were stratified by age, sex and season. Multiple comparisons were corrected using false discovery rate(FDR), and the robustness was evaluated through 10 sensitivity analyses. Results: A total of 1,160 individuals were included in this study, among whom 287 individuals(24.7%) involved the high-lethality suicide methods. Each 1 °C increase in daily mean ambient temperature was associated with a 2.2% reduction in the odds of selecting a high-lethality suicide method(OR=0.978, 95% CI: 0.964-0.993, P=0.004). After FDR correction, the temperature effects in the 30-59 age group and in autumn reached statistical significance; however, likelihood ratio tests for interactions revealed no significant effect modification(P>0.15). In the method-specific analyses, sharp instrument injuries(OR=1.021) and falls from height(OR=0.969) remained significant after FDR correction, and the direction of effects across methods was consistent with the lethality classification framework. The results were consistent across the 10 sensitivity analyses, with ORs remaining stable in the range of 0.976-0.985. Conclusion: Ambient temperature is negatively associated with the lethality of suicide methods, and rising temperatures are accompanied by a decline in the proportion of the high-lethality suicide methods. These results remain robust across multiple sensitivity analyses.