Lithium-ion (Li-ion) batteries have become the main power source of electric vehicle (EV) in recent years. In the event of crash, debris can penetrate the battery and cause short circuit, thermal runway or even explosion. Such instances put the passengers at risk. In this work, we have utilized numerical models to predict the deformation and failure response of a typical Li-ion battery module. A simplified finite element model for battery module that balances the model fidelity and computational cost was developed and validated with the test data. Based on this model, the solid steel protective cover plate for shock mitigation in the original module design was replaced by a sandwich structure with the same mass. Three different types of sandwich core were considered, namely, (1) hexagonal honeycomb, (2) metal foam, and (3) pyramidal lattice. The simulation results indicated that the pyramidal lattice core overperformed the others in terms of energy absorption and force transfer. Next, the numerical model of the battery module with aforementioned new sandwich cover plate was combined with machine learning (ML) techniques. By systematically adjusting the values of design variables, a design space was created, and the simulation was carried out for each design case to compute its energy absorption, initial peak force, and displacement at which initial peak force. Two machine learning algorithms, namely decision tree (DT) and artificial neural network (ANN) were applied to mine the simulation dataset. Decision tree modeling results identified the key design variables and their interrelationship. The ANN model served as a predictor of performance parameters based on the design variables. A comparison with the simulation results verified the accuracy and robustness of both ML models. The methods developed in this work have the potential to be applied in the safety design of energy storage systems.