In bioscience and health protection, data is expanding fast. In clinical information, precise research can assist early infection diagnosis, patients' social insurance, and community services. Prediction is an significant feature in the field of health care. In this work, we develop ML and deep learning algorithms for forecasting chronic illnesses in patients. Experiment with the modified prediction model from the supplied standard dataset. The goal of this paper is to predict chronic diseases in individual patients using machine learning methods such as K-nearest neighbor, decision tree, and deep learning with corrected linear activation function and Adam as an optimizer. When compared to many standard algorithms, the suggested system's accuracy improves. When compared to other algorithms, deep learning algorithms have a higher accuracy of around 96.7%. These methods are used to forecast chronic illnesses such as heart disease, breast cancer, and diabetes.