In order to control the seismic response of civil engineering structures more reasonably and conveniently, a fuzzy control algorithm based on BP neural network was proposed. The neural network was trained with the structural seismic dynamic response data to establish the structural analysis model, and the time-domain modal coordinates were taken as the controlled variables to reduce the order of the system, so that the number of fuzzy reasoning required to establish the modal fuzzy control rules was within the acceptable range, and the system energy minimum was taken as the control target to formulate the control rules. The fuzzy control numerical model of structural dynamic response was established to evaluate the damping effect of the proposed algorithm based on the calculated value of seismic dynamic response. The results showed that the trained BP neural network could accurately predict the seismic dynamic response of the structure and establish fuzzy control rules accordingly. Using mode fuzzy control only for the first mode of the structure could achieve satisfactory damping effect. When active mass driver(AMD) optimal control amplitude was used as the control domain of each floor, the damping effect of modal fuzzy control was different from it. A better damping effect could be obtained by increasing the control field.