Abstract: Traditional industrial structure analysis usually regards influence coefficient as a positive factor in choice of leading industry. However for industrial sectors in a recession，higher influence coefficient often means severe obstruction to the development of regional economy and awkward social problem. Moreover influence coefficient could not measure practical impact of a certain industry accurately due to the neglecting of scale of industry. Thus the factor of influence coefficient should be treated cautiously. New method is discussed to weight influence coefficient by proportion of added value of industrial sector in regional economy by which actual influence of industry on total economy could be measured more accurately. Base on it a modeling is introduced to analyze each sector of regional industry combined by both weighed influence coefficients and speed of development with the method of quadrants. Thus the modeling could measure the impact of both influence and development rate of industries. A case study of Xinjiang Autonomous Uygur Region is taken using 42 sector input-output table of Xinjiang. Results show that 5 industrial sectors including Extraction of Petroleum and Natural Gas and Raw Chemical Material and Chemical Products are high influence-high growth industries. These sectors have stronger driving effects to regional economy. 4 industrial sectors including agriculture and construction are high influence-low growth industries．These sectors have severe obstruction to regional economy and likely to cause series of social-economic problems. 12 industrial sectors including Manufacture of Telecom Equipment，Computer and Other Electronic Equipment are low influence-high growth industries. These industries are expected to become new leading industries. The new modeling of industrial analysis overcomes disadvantages of traditional methods of influence coefficients. The application in industrial instruction of Xinjiang gives conceive conclusions and valuable advises. It proves the modeling is effective and operable. Industrial construction is a complex system. The evaluation to industrial construction refers not only added values but also benefit per unit of output，labor payment，employment rate and so on. Advanced studies should be done to improve the modeling and therefore to measure efficient of industrial construction accurately．

Key words: weighted influence coefficient, input-output table, industrial structure, Xinjiang