Abstract:Innovation is an important engine for cities to enhance their core competitiveness in the knowledge age. The concept of innovation development has been deeply rooted in people’s minds, while its performance in China needs to be quantitatively evaluated. This paper uses the DEA-Malmquist index to measure the regional innovation development performance level, spatial structure, and change trend of 285 prefecture-level cities in China during the period from 2012 to 2016. The system clustering method is used to group areas with similar innovation development performance levels. Analyze the innovation development performance characteristics, problems and corresponding solutions in different cluster cities. Empirical research shows that regional innovation development performance in China is at the middle and upper level, but internal differences are large, and economically developed regions urgently need to solve resource allocation problems; the increase in total factor productivity is mainly due to the improvement of the overall level of technology, the next step what needs to be solved is how to gradually adjust the current scale so as to rationalize it; cities with similar performance levels but at different levels of economic development also have different ways to improve innovation development performance.