Sandwich structures are employed in many different fields including automobile, marine, and aircraft structures. However, debonding may take place at the core-face sheet interface, reducing the stiffness of the structure. Debonding may occur for a variety of reasons, including initial manufacturing faults, changes in service loads, tool drops, and foreign object impacts. It is critical to comprehend how debonding zones impact the vibration of sandwich structures because decreases in the natural frequencies (NF) could lead to a structure vibrating at resonance and lead to structural failure. This paper investigates the influence of debonding shapes and debonding locations on the free-vibration behavior of sandwich structures. Different sandwich structures that have varied debonding shapes at various locations are modeled using COMSOL MULTIPHYSICS. Debonding is modeled by using the CZM model. Validation studies were performed to validate the current study. After the validation study, free vibration analysis of all the sandwich structures was performed and the first six NF were obtained from the simulations. The results show the influence of the debonding shapes and debonding locations on the NF of the sandwich structures. From the results, it was observed that both the debonding shapes and debonding locations significantly change the NF of the sandwich structures. The debonding shapes cause a reduction and an increase depending on the debonding location. It was also revealed that both debonding shapes and debonding locations have a significant effect on the vibration behavior of sandwich structures. Using this method, the debonding shape and location, delamination shape, and location can be predicted using machine learning algorithms. This study includes free vibration analysis of sandwich structures with different debonding shapes and locations, and the results show that natural frequencies change depending on the debonding shapes and locations. This information can be implemented in machine learning for use in the field of damage detection and utilized to predict the shape and location of delamination in sandwich structures.