The invention discloses a particle filtering-based SLAM optimization method and relates to robot simultaneous positioning and map construction. The problem of failure task execution of a robot causedby noise interference received by the robot during task execution is solved. The particle filtering-based SLAM optimization method comprises the following steps of 1, defining a variable; 2, buildinga probability model according to the variable, wherein the probability model building comprises motion probability model building, sensor observation model building and grid map model building; and 3,optimizing an SLAM according to the particle filtering and probability model, wherein the step of optimizing the SLAM comprises the steps of automatically determining a position by the robot after the map is updated each time, and the automatic position determination of the robot comprises initialization, sampling, scanning, matching, weight calculation and sampling again. The particle filtering-based SLAM optimization method has the advantages that a true state of a target is obtained, and the failure task execution of the robot caused by noise interference received by the robot during taskexecution is prevented.