<p>Reliable tailings dam displacement prediction is crucial for structural health testing of tailings facilities, management decision-making and warning of dam failure risk. However, traditional tailings dam displacement prediction methods focus only on point prediction, which is far from meeting the requirements for risk monitoring and safety management of tailings dam systems. Considering the volatility and nonlinearity of dam behavior sequences, a hybrid interval prediction model incorporating complete ensemble empirical mode decomposition with adaptive noise, an intelligent optimization algorithm, an Elman neural network, and interval prediction is proposed for dam displacement interval prediction and analysis systems. This model can achieve pattern decomposition of nonlinear dam behavior sequences to determine the optimal feature mode and reduce modal mixing and reconstruction errors. The optimization algorithm is used to determine the optimal prediction weights and thresholds of the Elman neural network. Next, on the basis of the interval prediction principle, the upper and lower bounds of the interval prediction are obtained. Four-season displacement sequences were collected from a tailings dam in Anhui, China, and modeled to verify the effectiveness of the model. The simulation results show that the hybrid prediction model closely follows the actual curve trends and provides a superior prediction interval, exhibiting higher coverage and a narrower interval width than the other models. This model effectively considers the volatility and uncertainty in dam behavior sequences, ultimately enhancing the reliability and robustness of tailings dam displacement prediction.</p>