Since the outbreak of the financial crisis, the traditional way of financial supervision has been unsustainable, and countries around the world have carried out regulatory reforms. In this context, macroprudential policies have become the focus of attention of policymakers and scholars. Based on the BGGDSGE model, this paper first constructs the characterization index of systemic risk of the banking industry based on the endogenous bank bankruptcy mechanism, then introduces systemic risk factors into the classic welfare loss function, establishes macroprudential policy rules, and finally simulates the deduction transmission path and policy collocation through the impulse response function, and obtains the following conclusions: First, the macroprudential regulatory system with dynamic capital adequacy ratio and loan-to-value ratio (LTV) as the core should take output gap and credit volume as signal sources. Macroprudential monetary policy rules should be pegged to credit volume indicators, and capital asset prices are also suitable as reference variables; Second, macroprudential monetary policy and macroprudential supervision both play a role in maintaining financial stability, but as a standing policy, macroprudential supervision is better than macroprudential monetary policy; Third, identifying the drivers of economic cycle fluctuations is a fundamental issue for establishing macroprudential policy rules, enabling policies to be transmitted to the real economy more effectively, and playing a role in risk prevention and control. Fourth, when the two types of macroprudential policies are used jointly, the reasonable establishment of policy rule coefficients can enhance the positive spillover effect of policies, but the intensity of policy matching is not easy to control. In order to prevent the problem of effect interference caused by policy mismatching, only macroprudential supervision can be used during normal times, and when the risk is high, the two types of policies can be combined to achieve financial stability.