High blood sugar levels and serious consequences may occur from the prevalent chronic illness known as diabetes if the body’s insulin synthesis or utilization is disrupted. Predictive analytics in healthcare plays a crucial role in early detection. Current research developed predictive model for Pima Indian Diabetes Database using sophisticated deep learning (DL) techniques. Long short-term memory (LSTM), convolutional neural networks (CNN), and transfer learning are employed to analyze a comprehensive set of health metrics, including pedigree function of diabetes, body mass index (BMI), blood pressure (BP), skin thickness, insulin levels, age, glucose levels, and pregnancy history. The dataset is divided into training and testing subsets to evaluate and refine performance of model. Metrics that includes accuracy, precision, recall, and F1 score are assessed using a confusion matrix, enabling the selection of optimal algorithms for predicting diabetes. This research aims to facilitate accurate and timely healthcare interventions.