Abstract: In order to solve the modifiable areal unit problem in previous research, This study uses the point pattern analysis method with the fine-grained patent address space data of 2013 in Hunan Province, to analyzes the spatial pattern of innovation activities in Hunan Province and its influencing factors. The study finds that: There is heterogeneity in the spatial distribution of innovation activities in Hunan Province, showing a single-center “core-periphery” structure spreading out from Changsha, Zhuzhou and Xiangtan district and the best geographic scale for innovation spatial agglomeration is between 20km and 21km. After controlling the influence of innovation input factors, the GDP per square kilometer and the point distance from the road have significant impact on the intensity of innovation activity distribution. And because of the discontinuity of economic activity, the GDP per square kilometer has a differential impact on innovation activities. Lastly, we conclude that the policies to encourage the development of innovation should fully consider the imbalance of regional’s economic development. while incentives for independent innovation should be promoted in developed regions, road accessibility that increasing innovation spillovers in agglomeration should be emphasized in developing regions.