The most effective strategy to combat global disease spread is to prevent epidemics. As a result, develop a wearable safety and health monitoring device that utilizes the Internet of Things to provide authentic health testing and improve global health. However, because epidemics are nonlinear and dynamic systems, establishing preventive techniques remains a difficult task. IoT data integration is a major concern for disease prediction. To deal with this situation, a new classification and regression model (CART) has been presented to forecast data point labels. It overcomes scaling, outliers, and missing data information by traversing a binary tree according to the kernel to get the relevant labels. When the data are successfully integrated, there is a problem with noise and demographic bias in the prediction of the disease. Hence a novel Imbedded Least Square Support Regression (ILS-SVR) reduces random noise and removes drifts. In addition, to mitigate the demographic distortion of an unexpected regression problem, the theory by-law function (TBF) has been implemented. Furthermore, it helps to predict the epidemic disease better and for prevention strategy among the people, human mobility plays a significant concern. In order to tackle human mobility, a novel IoT-based global Preemptive scheme has been introduced to track the pattern and the harmful symptoms of infectious patients observed. This explores the possible role of K-NN machine learning techniques as a Hidden Markov Model (HMM). Thus automated alert systems prevent mortality, morbidity rate, and timely detection of epidemic diseases with high efficiency.