This paper details the rationale, structure, and potential applications of a novel screening tool for predicting substance use disorders (SUDs), highlighting how machine learning can revolutionize SUD risk assessment and inform targeted prevention and intervention strategies. SUDs pose a significant public health challenge, profoundly affecting individuals, their families, communities, and the healthcare system at large. Given these alarming trends, it is imperative to prioritize the early and precise identification of individuals at risk, enabling timely intervention and preventive measures. This study proposes the development of a novel screening tool that merges the well-established CRAFFT 2.1 questionnaire with genetic testing to formulate a comprehensive risk score enhanced by the predictive power of machine learning algorithms such as Random Forest (RF). By identifying complex patterns among genetic and behavioral data using machine learning, this study aims to overcome the limitations of traditional screening methods, which rely heavily on self-reported information and often fail to capture the intricate interplay of genetic predispositions and behavioral patterns in SUD risk. This comprehensive methodology aims to deliver a thorough risk assessment, enhance the precision of identifying at-risk individuals, and facilitate timely, personalized interventions tailored to the unique needs of each individual.