A novel nonlinear system identification method based on the Multi-dimensional Taylor Network is proposed in this paper. It provides a new method to identify the nonlinear system, which can describe the dynamic characteristics without prior knowledge and can realize the prediction of the nonlinear time series just with input-output data. Firstly, the structure of the Multi-dimensional Taylor Network is introduced. Then, a dynamic model based on the Multi-dimensional Taylor Network is established. Finally, this method is applied in construction safety monitoring to predict the vibration response of building structures. Results from the simulation example demonstrate the effectiveness and feasibility of the proposed method.