A beam structure in presence of crack subjected to both bending and axial load has been investigated for damage diagnosis to predict crack location and crack depth based on an optimized intelligent technique using genetic algorithms approach. Numerical with experimental investigations have been conducted in a free-free beam model to calculate natural frequencies. In this work, genetic algorithm as an optimized artificial intelligent technique has been used which employs first three relative modal parameters like natural frequencies as input data taken from numerical and experimental results to obtain optimized crack locations and crack depths as output parameters. This methodology has been found to be quite reliable for fault diagnosis by monitoring the possible variations in the relative natural frequencies of a free-free beam element in presence of crack. This paper also gives an insight into the deviation of the results of the proposed methodology from the experimental result.