Previous studies have reported the association between daily temperature and morbidity at city level, but small-area association of hourly temperature and acute morbidity is limited. We aimed to assess the association between hourly temperature and acute morbidity at the smallest administrative units (communities), and to explore the sources of regional heterogeneity. This case time series design was conducted using 145,678 non-accidental ambulance dispatch records in 328 communities in Fuzhou, China (2019-2024). Small-area associations between hourly temperatures and ambulance dispatches were assessed by quasi-Poisson regression with distributed lag nonlinear model. Multivariate meta-regressions used to explore modification of natural and social features. Unstandardized and standardized attributable number calculated for disease burden. We observed a U-shaped exposure-response curves of hourly temperature and risk of ambulance dispatches, with significantly adverse cold and heat effects referent at minimum risk temperature of 21 °C. The male and the elderly were more susceptible to non-optimal temperature. The unstandardized attributable numbers of non-accident ambulance dispatches due to cold and heat were 1,389 (95%eCI: 875, 1,748) and 590 (95%eCI: 319, 811), with corresponding standardized attribution number of 36 (95%eCI: 23, 46) and 15 (95%eCI: 8, 21). The standardized attributable number due to cold and heat were greater in the downtown and areas of higher proportions of the elderly and the higher population density, and lower NDVI level. High greenness mitigated heat-related adverse effects and temperature-related disease burden. The finding support the optimization of urbanized environment and public health strategies to reduce temperature-related disease burden under global warming.