The efficiency of a diagnosis system depends on the relevance of the information it can be retrieve from the diagnosed plant. The diagnostic system detects, localizes and assesses the damage which may be in the form of delimitation, crack, buckling etc. Optical sensors like Fiber Bragg Grating sensor may be used to determine the strain, temperature and other diagnostic parameters. In real world structural applications the numbers of sensors available are limited which thus requires the need for an efficient sensor optimization algorithm. This paper describes an approach in which a genetic algorithm is used to determine the optimum sensor positions for a diagnostic system. The fundamental issue regarding any diagnostic systems is which data should be processed and how the sensors should be placed. GAs being general purpose optimization algorithms is extremely suitable for such sensor placement problems. In this work the GA is tested using fitness function based on an assigned weight matrix and is developed while setting up the problem to simulate the presence of an effective damage detection system. For initial testing of the GA, we consider that there exists a neural network based identification system for a plate with 20 potential sensors locations.