Abstract: Analyzing the redundancy of investment in S&T innovation deeply, and having a good understand of technological innovation efficiency and its differences of 31 provinces, can help us better use technological innovation resource and improve efficiency. This paper constructs technological innovation efficiency index system considering labor input，capital investment and technical conditions basing on Cobb-Douglas production function. And measures the technological innovation efficiency of 2009-2013 S&T innovation employing CCR-DEA、BCC-DEA and Malmquist-DEA models. And this paper also divides 31 provinces into several types according to the input redundancy. The results show that in our research period, the quantity of fully effective provinces are increasing but still less than a half. The efficiency of S&T innovation based on technological progress has been increased within 5years. What’s more, only 8 provinces are areas which have no inputs redundancy. For the nation, the further reformation of S&T management policies and optimization of S&T resources allocation are important. 31 provinces should conduct different policies of S&T innovation with local characteristics basing on S&T resources and its redundancy.

Key words: S&T innovation efficiency, Malmquist-DEA, input redundancy, provincial difference