Objective Climate change has a significant impact on human health. This study aims to examine the relationship between humidity and respiratory related mortality in China. Methods We collected the daily meteorological data and respiratory mortality from 11 cities in Zhejiang Province during 2010 -2013. We applied distributed lag non-linear models (DLNM) with the indicator of attributable fraction (AF) to estimate the city-specific effect of humidity on respiratory mortality, and then used multivariate Meta-analytical model to pool the effect estimates in the 11 cities. Results We analyzed the total of 63 589 respiratory deaths during the study period,among which 6.16% (95% eCI: 1.97% -9.34% ) of mortality were attributable to non-optimum humidity. We observed a significant cumulative effect due to low humidity with an AF of 4.69% (95% eCI: 1.79% -6.94%) while that due to high humidity was 1.47% (95% eCI: - 1.17%-3.57%). The minimum mortality humidity (MMH) increased from 44% to 93% with a median of 88% which was corresponding to the percentile ranged from 80th to 95th percentiles in 8 cities. Both attributable fractions of high and low humidity were higher among males, youth and inland people in comparison with females, elderly and coastal people. Conclusions Most of the humidity-related mortality burden was attributable to the low humidity. This evidence has important implications for reducing the risk of extreme humidity for respiratory mortality.