This paper uses Dobson spectrometer total ozone data, Total Ozone Mapping Spectrometer (TOMS) data and radiosonde reports from Kunming, which is located in southwest China, from 1980 to 2008 to analyze the total ozone-climate relationship. The total ozone decadal long-term trend and abrupt change were studied using enhanced Dobson data whose missing data were amended by the TOMS data. Stepwise linear regression was used for the selection of the key factors that influence total ozone, including temperatures, geopotential heights, depressions of the dew point, wind velocities, and total solar radiation. The relationship between the selected factors and total ozone was analyzed using the methods of stepwise regression and partial least squares regression (PLSR). Results showed that although the PLSR method was slightly better and more reasonable to study the relationship than stepwise regression, while the two regression results were only slightly different. It was also suggested that local climate, especially local circulation and temperature, were important for the variations in total ozone, and the local climate could almost linearly explain 80% of the variance of total ozone. The relationship also indicated that the abrupt change of total ozone in the year 1994 may be related to abrupt local climate change.