Abstract:As the most active factor of productivity, scientific and technological progress is becoming the leading force for the leap of national competitiveness. Based on the theoretical framework of the regional innovation system, a data set of technological innovation indicators for the input and output of R D elements in Chinese cities from 2000 to 2019 was established. Identify innovation heterogeneity characteristics such as R D investment type, innovation process, and time lag from the knowledge production process. Using methods such as stochastic frontier analysis and data envelopment analysis, the technological innovation efficiency of Chinese cities is systematically measured. The relative effectiveness of the measured results is compared with various correlation coefficients, and the stochastic frontier results considering the process time lag are representative. The time series trend of innovation efficiency is tested by convergence analysis, and regional differences are calculated by Dagum Gini coefficient, the results show that China"s urban innovation efficiency has obvious dynamic convergence characteristics, but there are regional differences in the convergence trend, eastern innovation efficiency developing level is significantly higher than western and central regions, together with the agglomeration trend of innovation achievements in these regions, the dominant trend of innovation efficiency in high-level cities is also strengthening. Efficiency improvement is the basic way to promote the high-quality development of technology, we should continuously improve the regional innovation system, promote the smooth flow of R D factors, implement the promotion policy of regional equal allocation of innovation resources, and continuously improve the regional innovation efficiency.