A virtual maintenance system in a virtual environment can be used to simulate a real-world maintenance system. The efficiency of the simulation depends mainly on the assembly/disassembly task sequence. During simulation, path planning of mechanical parts becomes an important factor since it affects the overall efficiency of the maintenance system in terms of saving energy and time. Therefore, planners must consider the path-planning factors under constraints such as obstacles and the initial/final positions of the parts, as well as the assembly sequence such as number of gripper exchanges and direction changes. We propose a novel optimal assembly algorithm that considers the assembly sequence of mechanical parts and the path-planning factors for a virtual maintenance simulation system. The genetic algorithm is used to determine the optimal sequence of parts to minimize the numbers of gripper exchanges and direction changes, as well as find a repulsive force radius by using the potential field method to generate the shortest optimal distance for transferring each part during the assembly operation. By applying the proposed algorithm to a virtual maintenance system, users can be haptically guided to the optimized assembly solution during mechanical parts assembly operations.