The present research introduces an optimized framework for signature recognition. This framework incorporates three distinct methodologies: Loci method, Genetic Algorithm (GA), and Fuzzy Min Max Classifier (FMMC). The primary objective of this study is to achieve an enhanced recognition accuracy. The Loci method is a well-established approach used for pattern characterization, particularly suitable for patterns with strokes, especially in horizontal and vertical orientations, in this study, it is employed for signature recognition. However, a drawback of Loci method is its reliance on three critical hyperparameters: the degree of image division, the extent of zero padding and the restricted number of strokes. The choice of hyperparameters values significantly influences the efficiency of the recognition system. As a result, a GA is applied to ensure optimal hyperparameters values selection. This algorithm uses three phases with new strategies; selection, crossover, mutation; iteratively with the goal of maximizing the fitness function. The proposed fitness function is defined as the recognition rate calculated using FMMC. The effectiveness of the proposed methodology was tested and assessed through a comparative analysis with well-established models using an overly diversified signature database. The proposed model produced favorable outcomes compared to other approaches designed for the same task.