Large-scale civil infrastructures play a vital role in society, as they ensure smooth transportation and improve the quality of people's daily life. However, they are exposed to several continuous external dynamic actions such as wind loads, vehicular loads, and environmental changes. Interaction assessment between external actions and civil structures is becoming more challenging due to the rapid development of transportation. Data-driven models have lately emerged as a viable alternative to traditional model-based techniques. They provide different advantages: timely damage detection, prediction of structural behaviors, and suggestions for optimal maintenance strategies. This chapter aims to describe the advantages and characteristics of data-driven techniques to predict the dynamic behavior of civil structures through an artificial neural network (ANN). The applicability and effectiveness of the proposed approach are supported by the results achieved by processing the measurements coming from a monitoring system installed on a cable-stayed bridge (the Éric Tabarly Bridge in Nantes, France). Accelerations recorded by a network of 16 mono-axial accelerometers and Nantes Airport weather data acquired with the observation platform of the METAR (Meteorological Terminal Aviation Routine Weather Report) Station Network have been used as training to predict the structural response and to statistically characterize the behavior through a nonlinear autoregressive (NAR) prediction network. The performance has been evaluated through statistical analysis of the error between the measured and predicted values also related to both environmental conditions and the number of signals. The results show that the forecast network could be useful to detect the trigger of anomalies, hidden in the dynamic response of the bridge, at a low computational cost.