OBJECTIVES: To explore the association between weather conditions and hospital admissions for pneumonia in Shanghai. DESIGN: A time-series analysis was performed for a period of 4 years (January 2008-December 2011). A generalised additive model was used to calculate the relative risks. SETTING: Shanghai, China. PARTICIPANTS: All daily hospital admissions for pneumonia were obtained from the Shanghai health insurance system between 1 January 2008 and 31 December 2011 (n=99 403). RESULTS: The relationship between the mean temperature and pneumonia hospital admissions followed a V-shaped curve, with an optimum temperature (OT) at 13 degrees C. When the mean temperature was below the OT, a 1 degrees C decrease corresponded to a 4.88% (95% CI 2.71% to 7.09%) and 5.34% (95% CI 2.04% to 8.74%) increase in pneumonia hospital admissions in lag 4 using a single-day lag structure and lag 0-7 using a multiday lag structure. When the mean temperature >/=OT, no adverse effects from the temperature on pneumonia hospital admissions were found. The magnitude of the effects of temperature varied across gender and age groups. Hospitalisations for pneumonia increased by 15.99% (95% CI 0.06% to 34.46%) in the cold period. CONCLUSIONS: Cold temperature may be one of the important risk factors for pneumonia hospitalisations. Prevention programmes are needed to reduce the impact of cold temperature on pneumonia hospitalisations such as developing a weather warning system within a wide public health context.