This article presents an innovative method for predicting ship trajectories using ACoAtt–LSTM and AIS data to enhance maritime navigational safety. The study addresses the challenges of accurate and efficient ship trajectory prediction in complex marine environments. The proposed method integrates data encoding representation, an attribute correlation attention module (ACoAtt), and long short-term memory (LSTM) networks. The data encoding process transforms raw AIS data into a more efficient format, preserving key information while reducing complexity. The ACoAtt module employs a multi-head attention mechanism to capture complex relationships between dynamic ship attributes, such as speed and direction, thereby improving the model's understanding of implicit temporal patterns. The LSTM network is used to process time series data, enabling accurate and rapid predictions of future ship trajectories. The study evaluates the proposed model using a historical AIS dataset, demonstrating its superior performance in trajectory prediction accuracy and stability compared to other classic and advanced models. The model's effectiveness is validated through experimental results, showing that it outperforms traditional methods in handling short-term trajectory prediction tasks. The ACoAtt module enhances the model's ability to capture long-term dependencies and dynamic interactions among ship attributes, leading to more accurate predictions. The integration of data encoding, ACoAtt, and LSTM networks provides a robust framework for ship trajectory prediction, contributing to improved maritime safety and efficient navigation management. The study highlights the importance of advanced machine learning techniques in addressing the challenges of ship trajectory prediction, emphasizing the need for innovative data handling strategies and sophisticated temporal analysis methods. The proposed method offers a promising solution for enhancing maritime navigational safety by accurately predicting ship trajectories and mitigating collision risks.This article presents an innovative method for predicting ship trajectories using ACoAtt–LSTM and AIS data to enhance maritime navigational safety. The study addresses the challenges of accurate and efficient ship trajectory prediction in complex marine environments. The proposed method integrates data encoding representation, an attribute correlation attention module (ACoAtt), and long short-term memory (LSTM) networks. The data encoding process transforms raw AIS data into a more efficient format, preserving key information while reducing complexity. The ACoAtt module employs a multi-head attention mechanism to capture complex relationships between dynamic ship attributes, such as speed and direction, thereby improving the model's understanding of implicit temporal patterns. The LSTM network is used to process time series data, enabling accurate and rapid predictions of future ship trajectories. The study evaluates the proposed model using a historical AIS dataset, demonstrating its superior performance in trajectory prediction accuracy and stability compared to other classic and advanced models. The model's effectiveness is validated through experimental results, showing that it outperforms traditional methods in handling short-term trajectory prediction tasks. The ACoAtt module enhances the model's ability to capture long-term dependencies and dynamic interactions among ship attributes, leading to more accurate predictions. The integration of data encoding, ACoAtt, and LSTM networks provides a robust framework for ship trajectory prediction, contributing to improved maritime safety and efficient navigation management. The study highlights the importance of advanced machine learning techniques in addressing the challenges of ship trajectory prediction, emphasizing the need for innovative data handling strategies and sophisticated temporal analysis methods. The proposed method offers a promising solution for enhancing maritime navigational safety by accurately predicting ship trajectories and mitigating collision risks.