Phonetic deviations (dyslalia) are frequent speech disorders that lead most patients to speech therapy clinics. Nowadays one can find software able to support the treatment of phonetic deviations with several different approaches. However, none of the most used alternatives are easily accessible, are low cost and/or are directed towards individualization for an individual patient care. In this research we propose a Computer Aided Pronunciation Training (CAPT) system, which is low cost, easy to use, displays information directed to the patient's problem and, most importantly, uses computational intelligence techniques to make the therapy of phonetic deviations less expensive, less dependent on an audiologist and more efficient. The built prototype was evaluated by a group of audiologists from Federal University of Pernambuco (UFPE), who answered a questionnaire for reporting impressions on the general aspects of the proposal and also provided their opinions for possible improvements.