Background In the context of climate change, the health burden associated with non-optimal temperature is becoming increasingly substantial. Emergency ambulance calls (EACs) are a sensitive indicator of acute health effects. However, most previous studies have relied on daily mean temperature (MT) or a single thermal stress indicator, and it remains unclear whether different temperature indicators yield materially different estimates of the health burden attributable to non-optimal temperature. Objective To compare multiple temperature indicators in evaluating the attributable burden of EACs associated with non-optimal temperature in Dezhou City, Shandong Province. Methods Daily EACs records and corresponding meteorological data in Dezhou from 2014 to 2023 were collected. A time-stratified case-crossover design with conditional logistic regression, combined with distributed lag non-linear models (DLNM), was employed to examine the exposure-response relationships and population attributable burden of EACs associated with the percentiles of daily MT and 12 commonly used thermal stress indicators. Results During the study period, a total of 952989 EACs were recorded in Dezhou. For all temperature indicators, the cumulative exposure-response curves showed a V-shaped pattern. Among the attributable fraction (AF) estimates, only those derived from the universal thermal climate index for outdoor shaded space (UTCI3) and wet-bulb temperature (WBT) were significantly lower than the estimates based on MT (P < 0.05); no statistically significant differences were observed for the other indicators. The AF based on MT was 5.24% (95%CI: 4.34%, 6.16%), and the attributable burden was mainly driven by cold temperatures (AF=3.57%, 95%CI: 2.65%, 4.48%). Spatially, high heat-attributable EACs rates per 1000 population were primarily concentrated in the central and southern Dezhou, whereas areas with high cold-attributable EACs rates were more dispersed. The overall spatial pattern of EACs attributable to non-optimum temperatures was more consistent with the distribution of cold-related risks. Subgroup analysis showed that, across susceptible population groups, the AF estimates derived from most thermal stress indicators did not differ significantly from those based on MT. Conclusion In Dezhou, most thermal stress indicators produce attributable burden estimates for EACs associated with non-optimal temperature that are similar to those based on daily MT, and cold temperatures contribute more to the overall attributable burden. The spatial distribution of EACs attributable to non-optimal temperature is dispersed, whereas heat-attributable hotspots are concentrated in central and southern Dezhou.