This article viewing Fiscal and Tax [2018]70 on refunding uncredited VAT for specific industries as a quasi-natural experiment and selecting the data of A-share manufacturing quoted firms from the 3rd quarter of 2016 to the 3rd quarter of 2019, a total of 13 quarters of consolidated data as research samples, constructs a DID model to explore the effect of refunding uncredited VAT on manufacturing enterprises. The findings suggest that the profitability of the main business of enterprises that enjoy refunding uncredited VAT has significantly increased. The results of mechanism tests show: Refunding uncredited VAT improves the intensity of internal cash flow, reduces external financing cost of manufacturing enterprises, directly promotes enterprises to expand production scale, and improves the profitability of enterprises' main businesses. In the short term, the increase of industrial investment and R&D investment will increase the cost and reduce the profitability of the main business; industrial investment will promote the profitability of the main business of the enterprise in the long run, but the impact of R&D investment is uncertain. In addition, the results of heterogeneity analysis show: The policy has a higher incentive effect on enterprises with higher uncredited levels, non-state-owned enterprises and enterprises in central and western regions.