Nonuniform distribution characteristics of random points generated by Monte Carlo method in the workspace of parallel manipulators were discussed. Based on the density of distribution of the multiple random variable function, the relationship between Jacobian matrix and the probability density function of workspace points generated by Monte Carlo method was deduced, and the relationship between the probability density of the output space and the manipulability of the parallel manipulator was also constructed. The manipulability of a 2-DOF 5R parallel manipulator was analyzed based on the Monte Carlo method. According to the working mode of the robot, the input space was divided into four subspaces, and the inverse singularity occurred in the boundary among different subspaces was pointed out. Simulation experiment illustrated that the Monte Carlo method could work well for manipulability analysis of parallel manipulators.