The emerging and consummating of innovation measurement methods can be treated as the result of the improvement of innovation recognition and theory. Ever since 1950s, with the co-evolution of recognition, theory and measurement, there came out the “Linear Model of Innovation”, “Innovation System Approach”, and “Theory of Innovation Ecosystem”, and the measuring methods of “Innovation Input and Output”, “Composite Indicator”, “DEA” and “Data Mining”, etc. The former 3 can measure the innovation capacity of the target unit from different angles, and have been widely used in practice. However, the complex, systematic, networking, interacting features of innovation activity can hardly be reflected with these method. In the guidance of the Innovation Ecosystem Theory, Data mining based on data integration may be a good solution and a prospective trend. And combing different methods will improve the accuracy of measuring results.