With the continuous accumulation of dam deformation monitoring data and the continuous increase of deformation measuring points, it often takes a lot of time to predict all deformation measuring points, which is easy to cause the problem of untimely feedback. Therefore, the fuzzy C-means clustering algorithm (FCM) was introduced to partition the dam according to the similarity of deformation laws. The whale optimization algorithm (WOA) was used to optimize the parameters of long short-term memory neural network (LSTM), and a dam deformation prediction model based on FCM-WOA-LSTM was established. The measured deformation data of a concrete double-curvature arch dam was used as sample data for prediction, and the prediction results were compared with those of LSTM model and SVM model. The results show that the average absolute error (M_MAE), mean square error (M_MSE) and root mean square error (R_RMSE) of the prediction results of FCM-WOA-LSTM model are the smallest among the three models, and the three evaluation indexes of the fitting section are close to those of the prediction section, respectively. Compared with the existing models, the FCM-WOA-LSTM model has higher prediction accuracy and better applicability.