The prevention of the breast cancer through exams of mammas screening is the only way to reduce the women's mortality. The precocious detection propitiates the treatment of the disease in initial phase and a precise diagnosis can avoid unnecessary evasive exams. The objective of this work is to develop a fuzzy system to aid the medical diagnosis of the breast cancer, Computer Aided-Diagnosis (CAD). The classification of breast masses depends on the selection of characteristics that best represents the patterns. The acquisition of the data is a fundamental step for the system performance, because the quality of the input data is a limiting factor for its analysis. In this work was used Haralick features texture extracted from mamma images with masses and without masses and the Fuzzy model of Takagi-Sugeno to classify regions of interest (ROIs). Besides, the obtained results were compared with other classifiers and a comparative analysis is presented.