Background With the increasing frequency of extreme weather events due to global climate change, the impact of air temperature on stroke deaths is gaining increasing attention. Objective This study aimed to evaluate the impact of average daily temperature on stroke mortality in Pudong New Area. Methods Data were collected from Pudong New Area from January 1, 2005 to December 31, 2019, including deaths due to stroke (hemorrhagic stroke and ischemic stroke) , meteorological information (air temperature, relative humidity, atmospheric pressure) , atmospheric pollutants 〔particulate matter 10 (PM10) , sulfur dioxide (SO2) and nitrogen dioxide (NO2) 〕. We used a distributed lag non-linear model (DLNM) to evaluate the impact of average daily air temperature on stroke mortality, and assess the lagged and cumulative effects. Age- and education attainment-specific analyses were performed to identify the susceptible groups. Results During the period, a total number of 30 337 deaths from stroke were reported in Pudong New Area of Shanghai, and the deaths due to hemorrhagic stroke and ischemic stroke numbered 11 221 (36.99%) and 15 308 (50.46%) , respectively. The average daily deaths from stroke, hemorrhagic stroke and ischemic stroke were 5.54, 2.05 and 2.79, respectively. The average daily temperature, relative humidity and air pressure in Pudong New Area were 17.25 ℃, 73.75%, and 1 015.83 hPa, respectively. The average daily concentrations of PM10, SO2 and NO2 were 68.87 μg/m3, 28.82 μg/m3, and 44.68 μg/m3, respectively. Using the average median daily temperature (P50=18.2 ℃) as a reference, the strongest cumulative effects of low temperature (P5=3.0℃) were observed within 0-14 days lagged (Lag 0-14) for stroke mortality〔RR=1.34, 95%CI (1.15, 1.56) 〕 and hemorrhagic stroke mortality〔RR=1.61, 95%CI (1.25, 2.08) 〕. The impact of low temperature on stroke mortality and hemorrhagic stroke mortality was significantly lagged and cumulated, but its impact on ischemic stroke mortality was not significant. The strongest cumulative effects of high temperature (P95=30.1 ℃) on stroke mortality 〔RR=1.26, 95%CI (1.12, 1.41) 〕 and ischemic stroke 〔RR=1.64, 95%CI (1.39, 1.93) 〕 were observed within 0-1 day lagged (Lag 0-1) . High temperature imposed an acute effect on stroke mortality and ischemic stroke mortality, but produced insignificant effect on hemorrhagic stroke mortality. No significant impact of low temperature was observed on stroke mortality in population < 75 years old, but the impact on stroke mortality in population≥ 75 years old was strongest within Lag 0-14 d. The strongest impacts of low temperature on stroke mortality in population with primary education or below, and in population with secondary education or above were observed within Lag 0-21 d and Lag 0-14 d, respectively. No significant impact of high temperature was observed on stroke mortality in population < 75 years old and population with secondary education and above, but the impact was the strongest on stroke mortality in population ≥ 75 years old and the population with primary education or below within Lag 0-1 d. The impact of high temperature on stroke mortality in population with primary education or below was significantly higher than that in population with secondary education and above within Lag 0-3 d (P