Today's scenario of rapidly increasing antibacterial drug resistance is a major problem. So, to prevent from microbial infection and multi-drug resistance AMPS (antimicrobial peptides) has been highlighted in recent years. AMPS are a unique group of shorter to longer chain of molecules that can target and resist the bacterial infection directly. However, identifying AMPS by lab-experiments is time consuming and costly. Therefore, it is significant to develop computational tool for AMPS prediction. Though some AMPS prediction tools have been developed recently, their performances are not well enough to distinguish the AMPS from anticancer peptides and anti-diabetic peptides. In this systematic study, the selected 180 peptide's predictions are analyzed through the SVM (Support Vector Machine) machine learning method. In addition to SVM, Gaussian classifier is used in this research for optimizations. The linear SVM method shows the best model for the classification of AMPS. The best performance was shown in class of zero violations as compared with class of one, two, three violations. This study anticipated smaller chain dipeptides show high potency against antibacterial drug resistance and prevent bacterial infections.