Parallel robots have many useful characteristics, compared to linear robots like their greater structural rigidity, among others. The process of designing parallel robots take in consideration many variables such the workspace, the performance and the influence of the dimensions on these features, which make it an iterative procedure known as optimal design. This article presents the optimal design of a five-bar mechanism by the utilization of genetic algorithms. For the analysis, the robot is considered entirely asymmetrical which generalize the results to a more significant scope. Also, the objective function of the algorithm was the global performance index which considered the statistical behavior of the local indexes on the workspace.