Network dismantling aims to identify the optimal dismantling nodes through a specific strategy to fragment the network structure and is applied to various practical problems, such as optimizing the network resilience, dismantling harmful networks, and more. Given that entities in complex systems have diverse interactive relationships, typically modeled by multilayer networks, solving the dismantling problem on multilayer networks is crucial. This study uses genetic evolution theory to solve the dismantling problem on multilayer networks, and obtains the best dismantling nodes through the two evolutionary processes of crossover and mutation. Experiments on real-world and artificial multilayer networks show that the method proposed in this paper outperforms baseline methods in terms of dismantling effectiveness.