TODAY, SEX IDENTIFICATION IS CONSIDERED AS AN IMPORTANT TASK IN INFORMATION TECHNOLOGY APPLICATIONS. THIS PAPER CONCERNS SEX IDENTIFICATION USING Support Vector Machine (SVM). RBF AND POLYNOMIAL AS TWO KERNEL FUNCTIONS WERE STUDIED. IT WAS OBSERVED THAT RBF KERNEL OUTPERFORMS THE POLYNOMIAL KERNEL FUNCTION. LPCC AND MFCC CEPSTRAL COEFFICIENTS AND THEIR FIRST DERIVATIVES WERE ALSO EVALUATED. THEY BOTH SEEM TO BE GOOD FEATURES FOR SEX IDENTIFICATION, BUT MFCC COEFFICIENTS WERE SHOWN TO RESULT A BETTER PERFORMANCE THAN LPCCS. ADDING FEATURE DERIVATIVES TO FEATURES VectorS WAS ALSO SHOWN TO IMPROVE THE SEX IDENTIFICATION PERFORMANCE.