A Closed-Loop CL system or Artificial Pancreas AP is a system developed to help those with Type 1 diabetes mellitus T1DM maintain their blood glucose level. In this research, we aimed to improve the effectiveness of the AP by predicting glucose using Deep Learning DL algorithms, specifically the Long Short-Term Memory LSTM algorithm. The trained LSTM model achieved between 60% - 90% accuracy on the training set and 80% on the test set. The model had good accuracy and recall for both groups, indicating its ability to quantify the adjustment dose of insulin and then pump it out through the AP. In the future, we hope to have a fully automated AP that can be implemented through an application allowing diabetic patients to take pictures of their meals within the app. The amount of carbohydrates in the meal will be automatically calculated through scanning, and this information will be communicated directly with the AP.