Cleft lip and palate represent the most prevalent craniofacial deformities in the global population. This study aims to investigate the prevalence of cleft lip and palate in the state of Paraíba, Brazil. This is an exploratory, retrospective and epidemiological study, with a quantitative approach. The research was conducted at the Lauro Wanderley University Hospital (UFPB) in two stages: the first consisted of collecting data from the medical records of patients treated by the hospital's Cleft Lip and Palate Service. In the second stage, statistical inference techniques were applied, such as Logistic Regression and machine learning Neural Networks, to predict variable patterns and profile the patients treated. In total, 2,638 patients were registered, of which 870 started care late, with the first consultation taking place at five years of age or older. This data points to a possible failure in the implementation of prevention policies in primary care. A higher concentration of cases was observed in the mesoregions of Mata Paraibana and Sertão Paraibano, with transforamen cleft affecting the most structures and being the most prevalent, and consultations being carried out late, compromising several functions associated with this condition. The logistic model indicated that the covariates sex, diagnosis and mesoregion are significant at the α=0.05 level as predictors of late access to the CLP reference service in the state of Paraíba. The late diagnosis rate of CLP in the state was 32,98%, with approximately 10.5% being over 20 years old, highlighting the urgent need for secondary prevention and early diagnosis campaigns, particularly in the context of oral health care.