As an important institutional reform in the Chinese capital market, the registration system reform aims to improve the disclosure quality of public companies. Using the exogenous shock from the registration system reform as a quasi-natural experiment, we probe the effect of the registration system on disclosure in two stages: IPO admission and post-IPO supervision. We find that the registration system reform significantly reduces the performance deterioration of "incremental" companies in the IPO admission stage and improves the disclosure quality of "existing" companies in the post-IPO supervision stage. In addition, curbing earnings manipulation is a possible channel the registration system reform affects disclosure. Furthermore, these effects are more pronounced in firms with lower internal governance levels and less external media attention. Collectively, these findings respond to the theoretical controversy over whether the registration system will improve or worsen disclosure quality and confirm the reform has positive effects.