In this paper, the architectural parameters optimization of a three-prismatic-universal-universal (3-PUU) parallel kinematic machine (PKM) with three translational degree-of-freedom (DOF) is performed using the efficient particle swarm optimization (PSO) to achieve the optimum stiffness characteristics. Based on the stiffness matrix derived previously, the minimum stiffness over a cubic usable workspace is considered as a performance index since the manipulation accuracy of the PKM is dependent on the minimum stiffness throughout the workspace. To illustrate the efficiency of the PSO approach, both the traditional direct search method and the genetic algorithm (GA) are applied to the optimization as a comparison. The simulation results show that the PSO is the best method for the optimization, and the results are valuable in designing a 3-PUU PKM for machine tool applications.