Although steroids such as corticosteroids and anabolic androgenic steroids (AAS) are frequently used in sports and medical therapies, they can seriously endanger the health of vital organs including the heart, kidney and lungs. Long term use of steroids has been associated with respiratory risks such as fibrosis and impaired lung function compromised renal dysfunction including nephrotoxicity and chronic kidney disease and cardiovascular problems like hypertension and myocardial infarction. This study predicts steroid induced organ damage using machine learning approaches particularly Random Forest and XGBoost. To find risk factors adverse effects dependent dose and early warning indicators, clinical data and patient biomarkers were examined. Early intervention measures were supported by the excellent accuracy of the models in identifying individuals at risk. Explainability of SHAP strategies were used to address issues such as data imbalance and interpretability of the models. The results highlight the need for regulatory measures and medical recommendations personalized health risks to mitigate steroid related health risks. Future research should integrated real time patient monitoring and genetic data to improve predictive accuracy. This study contributes to precision medicine by improving early diagnosis, prevention and patient outcomes of steroid induced organ damage.