The current study assesses the discriminatory power of formants of diphthongs /au, ai, oi/ in Malay for speaker comparison casework by using discriminant analysis. The formant dynamics of the diphthongs from 10 speakers are analysed, to measure the full trajectories of F1, F2 and F3 at 9 x 10% time intervals. The results show that diphthongs in Malay have the discriminatory power in distinguishing speaker. In reading text speech style the accuracy rate is between 77.9-81.8% and 18.9-36.1% in spontaneous speech style, from same speakers. The results also show there is consistency of accuracy rate of 75.8-85.8% within speaker from two different styles. In this current study, diphthong /ai/ in reading text speech and diphthong /oi/ in spontaneous speech style are the best variables used to discriminate speakers. Each diphthong has different best predictor; F3 for diphthong /oi/, F2 for diphthong /ai/ and F1 for diphthong /au/ but the best is by using all predictors of F1, F2 and F3. These results show promising potential of the Malay diphthongs to be used in forensic speaker comparison cases.