In the current research numerical method and hybrid artificial intelligent (AI) technique are adopted for prediction of cracks in a cantilever composite beam structure containing transverse non-propagating cracks. The presence of cracks in the dynamic structures like beam is a serious threat to the performance as well as life of structures and it affects the vibration signatures e.g. natural frequencies and mode shapes. The graphite fiber reinforced polyimide composite engaged in the present investigation due to its advantageous features, such as higher damage tolerance capability, improved fatigue resistance, high stiffness and strength-to-weight ratios compared with the metallic structures. The Numerical analysis has been completed by using commercially available analysis and simulation software package ANSYS to catch the relation between the change in natural frequencies and mode shapes for the cracked and un-cracked composite beam. Which consequently used, to the design of smart system based on hybrid AI technique for prediction of crack depths and locations following inverse techniques. The calculated modal frequencies, mode shapes, relative crack locations and relative crack depths are used to design the fuzzy neural model. The measured vibration signatures are used as inputs to the fuzzy segment of the hybrid model and initial relative crack depths and initial crack locations are the output parameters of fuzzy model. The first three relative natural frequencies, first three average relative mode shape difference and the output from the fuzzy model are used as inputs to the neural segment of the hybrid model and final crack depths and locations are the output parameters. The measured vibration signatures are used to formulate a series of fuzzy rules and training patterns for the fuzzy and neural model. Finally, the validation of the proposed method is carried out dynamically by means of experimental results from the developed experimental setup.