<p>Cold spells are intensively studied for their associations with mortality, but evidence is limited in low-latitude plateau regions. The purpose of this study was to investigate the impacts of cold spells on the mortality risks in Yunnan Province. Mortality and meteorological data were collected from 129 counties in Yunnan between 2014 and 2020 and summarized into 16 administrative regions as the research units. A quasi-Poisson-based distributed lag nonlinear model (DLNM) was used to assess the impacts of cold spells on mortality in cold seasons (October to March). Random-effects meta-analysis was applied to compute the pooled effects, and spatial heterogeneity and population vulnerability were evaluated by adjusting for regional characteristics and stratified analysis. The best model fit was obtained when cold spells were defined as the daily average temperature below the 10th percentile for 3 consecutive days. The pooled relative risks (RR) of non-accidental mortality for lag 0 days, lag 0–7 days, lag 0–14 days, and lag 0–21 days were 1.01 (95% CI: 1.01–1.02), 1.08 (1.04–1.12), 1.11 (1.06–1.17), and 1.13 (1.06–1.20), respectively. Higher cold spell effects were observed for cardiovascular and respiratory disease mortality, with RRs of 1.14 (1.06–1.23) and 1.22 (1.13–1.32), respectively. Populations in the more easterly regions of Yunnan, or individuals aged ≥ 75 years, exhibited higher risks of non-accidental mortality Cold spells increased mortality risk, with the effect modified by geography, age, and lag days. These findings indicate the necessity to develop regional prevention strategies and early warning systems.</p>