A low-carbon economy is an essential way to tackle global climate change.The essence of a low-carbon economy is to increase carbon productivity.Carbon productivity is often affected by impact factors of regional low-carbon development,including global variables and local variables.We need to consider these together in order to establish a precise model.Here we determined factors influencing carbon productivity in Chinese provinces based on spatial autocorrelation method.We found that the industrial structure is a global variable and that energy structure,technological progress and labor productivity are local variables.Four impact factors and regression parameter values at the end of 11th Five-Year Plan and 12th Five-Year Plan were estimated and analyzed based on mixed geographically weighted regression.We found that the structure of energy(power ratio)has a negative effect on carbon productivity,and industrial structure(service industry),technology(patent authorization number)and labor productivity (employees of the industrial added value)has a positive effect on carbon productivity.Judging from the estimated value of regression parameters,the influence degree of industrial structure occupies a leading position,followed by energy structure and technological progress,and finally for labor productivity.The impact of industrial structure on carbon productivity is increasing.The negative impact of energy structure on carbon productivity showed obvious decline from south to north spatially,while the positive impact of technical progress and labor productivity showed obvious declines from north to south.From the end of 11th Five-Year Plan to the end of the 12th Five-Year Plan,the impact of energy structure and labor productivity on carbon productivity decreased,and the impact of technological progress increased.Some policy suggestions are discussed.