Beijing and Shanghai are major sci-tech innovation hub cities in China. Based on urban ZIP code spatial database and R&D-intensive manufacturing firm data, this article explores the spatial distribution and influence factors of R&D-intensive manufacturing industry by using ArcGIS and Stata, and then reveals the difference of the location distribution model of R&D-intensive manufacturing industry between Beijing and Shanghai. The results show that Beijing R&D-intensive manufacturing industry tended to distribute mainly in the northwest inside the Sixth Ring and displayed the centralized pattern of Zhongguancun-Changping, Fengtai, Yizhuang and Wangjing-Jiuxianqiao as spatial aggregated hot zones, while Shanghai R&D-intensive manufacturing industry was characterized by the coexistence of concentration and diffusion, distributing in each direction inside the suburban loop line and taking Caohejing, Zhangjiang, Jinqiao, Meilong and Wujiaochang as spatial aggregated hot zones. Meanwhile, both Beijing and Shanghai R&D-intensive manufacturing industry had significant agglomeration effects, the difference is that Beijing took Zhongguancun-Shangdi as the single agglomerate center, while Shanghai took Caohejing and Zhangjiang as the twin symbiotic agglomeration pattern. The spatial distribution of different industrial types was unanimous with the general characteristics, but showing a certain specific. The high-technology firms were mostly located in the main urban areas, and their hotspots distribution shrunk back to the main urban areas and suburbs. Conversely, the other traditional manufacturing firms which also belong to R&D-intensive manufacturing industry were scattered in the outskirts, with the hotspots stretched out to the remote suburbs. Development zone, transportation accessibility and path dependence all had an important impact on the choice of location of firms in Beijing and Shanghai, and in addition the location of Shanghai firms was also affected by the positive impact of suburbanization and the location of research institutions and the old industrial areas, while suburbanization have negative impacts on the location of Beijing firms. As a result, the R&D-intensive manufacturing industry distribution in Beijing was a compact-central model with the development zone as the single industrial spatial carrier, while in Shanghai was a discrete-suburbanize model with development zones, research institutes and old industrial areas as multiple industrial space carriers. This study just used cross-sectional data without a time-series, so it could not analyze the evolution trend of the R&D-intensive manufacturing industry distribution and its formation mechanism. Also, the data possibly had survivorship bias. More importantly, the difference of economic and social benefits between different R&D-intensive manufacturing industry spatial distribution patterns needs to be discussed further.