Many engineering design problems involve a combination of both continuous and discrete variables. However, the number of studies scarcely exceeds a few on mixed-variable problems. In this research Particle Swarm Optimization (PSO) algorithm is employed to solve mixed-variable nonlinear problems. PSO is an efficient method of dealing with nonlinear and non-convex optimization problems. In this paper, it will be shown that PSO is one of the best optimization algorithms for solving mixed-variable nonlinear problems. Some changes are performed in the convergence criterion of PSO to reduce computational costs. Two different types of PSO methods are employed in order to find the one which is more suitable for adoption in this approach. Then, several practical mechanical design problems are solved by this method. Numerical results show noticeable improvements in the results in different aspects.