Contemporary speaker recognition systems are prone to errors due reverberation when operated in closed environments. The project focuses on mitigating the effects of reverberation on Automatic Speaker recognition system. Implemented in MATLAB, the system uses Gaussian Mixture Model and Maximum Likelihood Criteria for feature (MFCC) matching and exponential envelope removal for reverberation mitigation. An average improvement of 16% in recognition rate was observed. The project is to be further improved for blind mitigation.