In this paper, In this paper, we investigated the parameters that can determine whether or not to drink by comparing the voice characteristics before and after drinking using male and female speech analysis. The proposed method correctly extracts and categorizes characteristics such as changing rhythm and vocal state, and proceeded to judge drinking. Another method is to determine whether alcohol is consumed by using slope values of formants before and after drinking. The method of determining drinking by formant inclination is based on a typical frequency spectrum type. When the alcohol is consumed, the frequency spectrum form of the voice signal becomes flat due to high frequency components due to the non-sounding phenomenon. Using this method, speech signals were applied before and after drinking of male and female with speech signals. As a result, it was found that the shape of the drinking posterior formant was flattened irrespective of gender or gender. As a result, it can be said that it is effective to analyze and apply the pattern of the voice signal before and after drinking without determining the weight or classification of sex when judging whether or not to use the speech signal.