Integrity monitoring systems are intended to assess the reliability of the information provided by the navigation systems. Advanced Receiver Autonomous Integrity Monitoring (ARAIM) is a primary method for integrity monitoring of multiple satellite failures. The allocation of continuity risk is associated with the threshold of fault detection. High fault detection thresholds may increase the probability of hazardous misleading information, leading to a high integrity risk. However, existing ARAIM integrity monitoring methods often adopt an equal allocation of continuity risks, which leads to redundancy in the allocation of false alarm probabilities. This problem results in a conservative estimation of integrity, thereby reducing the availability of the method. This paper proposes an ARAIM optimization method based on Bayesian Optimization. In this method, the minimum protection level is taken as the objective function of the Bayesian Optimization. The continuity risks allocated to different fault modes are used as parameters to be optimized in the Bayesian Optimization. These parameters are normalized to connect with the continuity risk allocation. Constructing the surrogate model based on a Gaussian process and appropriately selecting the acquisition function can help optimize the allocation of continuity risks. Optimized continuity risk allocation improves the selection of fault detection thresholds for test statistics to achieve the optimal protection level. This method is implemented and validated, and the results are compared with those of traditional methods. The results indicate that the continuity risk allocation strategy based on the Bayesian Optimization algorithm optimizes the protection level and improves the global availability of ARAIM.