In order to clarify the impact mechanism of tax policy optimization on corporate innovation model, the VAT refund policy was taken as a quasi-natural experiment, and based on the panel data of China’s Shanghai and Shenzhen A-share listed companies from 2012 to 2022, the impact of VAT refund on corporate exploratory innovation preference was empirically analyzed. The results show that: the VAT refund has significantly enhanced the enterprise’s preference for exploratory innovation and assisted the high-quality development of enterprises, which has been firmly established under multiple robustness tests. Mechanism analysis and heterogeneity discussion show that VAT refund promotes firms’ exploratory innovation preferences through two channels: enhancing R&D intensity and improving operating conditions. Among them, non-state-owned enterprises, young enterprises and enterprises in high-tech industries are relatively more sensitive to policy effects. Re-verification from the perspective of technology merger and acquisition shows that the VAT refund policy will significantly enhance the exploratory innovation preference of enterprises after technology merger and acquisition, indicating that technology merger and acquisition is an effective path to optimize corporate innovation models. The research conclusions can provide reference for tax policy formulation, precise policy implementation and creating a good corporate innovation environment.