This article describes an experiment using fuzzy logic for recognition of crackles in respiratory sounds.It presents the theory and the separation of crackles from vesicular sounds (VS).Some articles which use similar techniques to the experiment are related and compared.The experiment uses wavelet transform (WT) to separate crackles which are used as training datato an adaptive neuro-fuzzy inference system (ANFIS). The generated fuzzy inference system (FIS) separated crackles from real lung sounds, and obtained root mean square errors(RMSE) varying from 0.0156 to 0.0349.