Trusses are important structures that are commonly used in bridges, roofs, towers, and other structures. These structures are often prone to damages during their service life due to various reasons. Structural health monitoring (SHM) is extremely helpful for detecting these damages, thereby enabling corrective measures on time. Acoustic emission (AE) is a sensor-based nondestructive technique that performs real time SHM without disturbing the structural integrity. A key issue of AE technique in SHM is the selection of optimal number and location of sensors, as it is impractical and uneconomical to place large number of sensors on the structure. In this paper an attempt is made to find the optimal location of sensors using AE technique for a laboratory scale truss model. To achieve this goal, support vector machine (SVM), a machine learning (ML) algorithm, is utilized to find the most preferable joint(s) of the truss where the sensors can be placed for identifying the damaged member. In this study, damage, which is actually an AE source, is simulated by pencil lead break (PLB) and identification of damaged member is treated as a classification problem. Various signal features are extracted from each signal and used as input to train the SVM model and the assigned member is taken as output. The trained model is then tested with another set of experimental data. The results are found to be encouraging in terms of accurately classifying the damage members and the approach is found to be suitable for determining optimal sensor locations (OSL).