Integrity monitoring of precise point positioning (PPP) can provide tightly guaranteed absolute position error bounds for safety-critical applications. However, complex local environment makes PPP integrity monitoring much more challenging, such as urban canyons. Significant prone multipaths and low observation redundancy are main difficulties to the accuracy and the reliability of PPP. Therefore, we proposed a solution separation-based integrity monitoring algorithm, which is based on a single and dual frequency-mixed undifferenced and uncombined PPP model considering compensation for the multipath error distortion by Gaussian overbounding. Both the static and the kinematic data are utilized to test the proposed algorithm. The results show that the proposed algorithm can produce adequate protection level in horizontal and vertical directions. Furthermore, the proposed algorithm can obtain smoother protection level and positioning error under the dynamic local environment, and effectively suppress the misleading information.