Since the early 2000s, R&D expenditure in China has increased rapidly, with the country having the third highest R&D expenditure, next only to the OECD and USA, and having surpassed Japan since 2009. Furthermore, policies in China in recent years have emphasized the need to conduct more basic and applied R&D in order to overcome technological bottlenecks and risks in access to advanced technologies when faced with international geopolitical tension. However, in the meantime, the trend in measured total factor productivity (TFP) in China has been downward since 2010. A lack of TFP growth despite the significant investment in R&D raises the question of the impact of R&D expenditure on productivity growth in China. Therefore, this study aims to investigate this issue and estimate the effects of three types of R&D stocks (basic, applied and experimental R&D stock) on TFP in China, using newly constructed provincial panel data in China from 1998 to 2018. Various empirical models and control variables are adopted to take into account non-stationarity and spatial spill-over of the provincial R&D stock values over time. The analysis results are robust to various specifications and reveal a significant positive effect of overall R&D and experimental R&D on TFP in China, but basic R&D exerts no significant results, and the effects of applied R&D are mixed across specifications. Further analyses using the 1991–2018 national data demonstrated largely consistent results. These results suggest that experimental R&D has been crucial for enhancing TFP growth in China during the decades investigated, but evidence of basic and applied R&D driving TFP growth in China is lacking.