A deep learning model based on 12-lead ECG was developed to predict left ventricular diastolic function and increased filling pressure using echocardiographic data. The model was trained, validated, and tested on 98,736, 21,963, and 98,763 patients, respectively, with ECG and echocardiographic assessments within 14 days. It was also tested on 55,248 patients with indeterminate diastolic function. The model demonstrated high accuracy in detecting increased filling pressure (AUC 0.911) and diastolic dysfunction grades (AUCs 0.847, 0.911, and 0.943 for grades ≥1, ≥2, and 3, respectively). The model showed similar prognostic performance to echocardiography, with higher mortality in patients predicted to have increased filling pressure. The AI-ECG model identified increased filling pressure and diastolic function grades with good prognostic value, making it a promising tool for detecting diseases associated with diastolic dysfunction and increased filling pressure. The study highlights the potential of AI-ECG as a simple and effective method for assessing diastolic function and filling pressure, complementing traditional echocardiography. The model was developed using a convolutional neural network and showed strong performance in predicting diastolic function and filling pressure. The study also discusses the limitations of the model, including the use of echocardiographic data as a reference and the need for further validation in diverse populations. The results suggest that AI-ECG could be a valuable tool for diagnosing and managing heart failure with preserved ejection fraction and other cardiovascular conditions.A deep learning model based on 12-lead ECG was developed to predict left ventricular diastolic function and increased filling pressure using echocardiographic data. The model was trained, validated, and tested on 98,736, 21,963, and 98,763 patients, respectively, with ECG and echocardiographic assessments within 14 days. It was also tested on 55,248 patients with indeterminate diastolic function. The model demonstrated high accuracy in detecting increased filling pressure (AUC 0.911) and diastolic dysfunction grades (AUCs 0.847, 0.911, and 0.943 for grades ≥1, ≥2, and 3, respectively). The model showed similar prognostic performance to echocardiography, with higher mortality in patients predicted to have increased filling pressure. The AI-ECG model identified increased filling pressure and diastolic function grades with good prognostic value, making it a promising tool for detecting diseases associated with diastolic dysfunction and increased filling pressure. The study highlights the potential of AI-ECG as a simple and effective method for assessing diastolic function and filling pressure, complementing traditional echocardiography. The model was developed using a convolutional neural network and showed strong performance in predicting diastolic function and filling pressure. The study also discusses the limitations of the model, including the use of echocardiographic data as a reference and the need for further validation in diverse populations. The results suggest that AI-ECG could be a valuable tool for diagnosing and managing heart failure with preserved ejection fraction and other cardiovascular conditions.