Based on the observational data of Shaanxi Province from 1961 to 2002, after minimizing the possible biases caused by urban heat,the first eigenvector values of maximum and minimum temperature were obtained by the empirical orthogonal function method,which are positive in four seasons and whose variance contributions are over 60 percent.The temporal coefficients of the first eigenvectors can reflect the variation of maximum and minimum temperature.According to these coefficients,the spatial and temporal distributions of trends for maximum and minimum temperature were studied.The results show both maximum and minimum temperature show an increasing trend after 1984 in all seasons.These trends were centered around the central and western parts of Guanzhong and Weibei.Except in autumn,an increasing trend for temperature daily range is found in other seasons over the recent ten years,which was possibly caused by the decreasing of cloudy quantity.