PURPOSE: to analyze the behavior of the Path Course, Level of Complexity and acquisition of phonemes, according to the classification of severity of phonological disorders (PD) through the Linguistic Fuzzy Model (BRANCALIONI, 2010). Also, to verify if the classification of severity by means of the Model is able to differentiate quantitatively the degrees or classes of sounds and distinctive features. METHOD: the sample was made up of the phonological assessment pre-treatment with 204 DF subjects of both genders and aged between four years and two months old to eight years and two months old. All subjects were assessed for phonological evaluation of understanding and expressing language, evaluation of the stomatognathic system, repetitive articulatory examination and audiological assessment. Contrastive and distinctive features analyses were performed. The severity of 204 deviant phonological systems was classified according to the model. The corresponding values of the variables were tabulated and analyzed statistically by Tukey test and Qui-Square Association complemented by analysis of adjusted residuals, considering a significance level of 5% (p <0.05). RESULTS: it was verified that the variables of Path Course, Level of Complexity and Acquisition of Phonemes decrease as the severity increases; these relationships are statistically significant. In addition, it was observed that the number of altered sound class and distinctive substitutions sound increase according to the severity; these relationships are statistically significant. CONCLUSION: the model is able to differentiate the degrees according to the variables, bringing important information according to different degrees.