Abstract: Industry-university collaborative innovation is a kind of innovation activity participated by heterogeneous organizations. How multi-dimensions of proximity between heterogeneous organizations affect industry-university collaborative innovation performance is a question deserve further exploration. Based on the perspective of proximity, uses joint-patent data provided by the State Intellectual Property Office patent retrieval database, adopts the method of negative binomial regression analysis, this study empirically analyzes the impact of multi-dimensional proximity on industry-university collaborative innovation performance, with a specific focus on the effects of institutional proximity and technological proximity, especially concern the interactive effects of geographical proximity and technological proximity. The results show that the promoting effect of institutional proximity is not significant; the effect of technological proximity on industry-university collaborative innovation performance takes an inverted-U shape, and this inverted-U shape relationship is negatively regulated by geographical proximity. Our results also suggest that geographical distance does not negatively affect the performance of industry-university collaborative innovation.