Sea-crossing bridges in China are mainly built in nearshore island area where wave condition varies spatially. The accurate estimation of the wave height in the bridge site is of great significance for bridge design and construction organization. An artificial neural network (ANN) estimation model of wave height in nearshore island area was developed based on open ocean environmental forecasting data. Pingtan Strait sea-crossing bridge site was selected as the research object. The BP neural network commonly used in the ANN was adopted to train the data provided by the open ocean forecasting station and the measured wave height data in the bridge site area. In order to verify the feasibility of the model, the wave height in the bridge site for 80 consecutive days was estimated. By comparing the results of previous model and the measured data, it is found that the trend of the estimation and the measured value is generally consistent. The root mean square error satisfies the prediction requirements and the ideal prediction effect is obtained. The research showed that the proposed ANN estimation model can use the open ocean forecasting information to effectively estimate the wave height of the nearshore island area for sea-crossing bridge with a relatively simple modeling process.