One of the most serious and expensive effects of diabetes is diabetic foot ulcers (DFUs), which can frequently result in infection, amputation, and a markedly reduced quality of life. For prompt comprehensive intervention and better patient outcomes, it is essential to estimate the risk of DFUs and comprehend their survival periods. In order to forecast the likelihood of developing DFU and the duration until ulcer onset and recurrence, this study offers a predictive model that incorporates machine learning approaches. Our method makes use of a hybrid machine learning model that blends Deep Neural Networks (DNN) with Random Forest (RF). The RF component makes use of its proficiency with organized, tabular data from clinical assessments and patient self-assessment instruments, while DNN is utilized to extract complex, nonlinear correlations and patterns from the identical dataset, augmenting the predictive power overall. Logit outputs from both models undergo adaptive fusion before logistic regression refines the final risk prediction. In order to facilitate the prediction of ulcer-free survival periods and the likelihood of ulcer recurrence, the research also uses Random Survival Forests (RSF) to model the time-to-event data. In order to classify patients into short-term, medium-term and long-term risk categories and enable customized treatment plans and monitoring schedules, time-to-event analysis. This project’s main goal is to evaluate the risk of DFU onset in diabetic patients and produce a reliable risk score that physicians can use to identify high-risk patients and refer them for early preventative care.