The growing prevalence of diabetes among people worldwide has caused the medical industry to explore alternate approaches for enhancing their healthcare innovations. Researchers are conducting a recent study in the fields of artificial intelligence to develop precise and efficient methods for diabetes detection. The present study extensively examines and analyzes the effects of current ML and DL methodologies on diabetes diagnosis and classification. Within this paper, we unveil a framework that harnesses advanced machine learning (ML) and deep learning (DL) algorithms for the precise classification of individuals afflicted with diabetes. The proposed framework utilizes fine-tuning of artificial neural network (ANN) models to obtain optimal results. We have performed an analysis of the proposed framework on two different diabetes datasets. In terms of accuracy, the proposed framework performs at 90%, which is higher than other algorithms.