A physiological “digital twin” model integrating both fast (BAC, BrAC) and slow (EtG, EtS, UAC) alcohol biomarkers was developed to improve reconstruction of drinking scenarios. By mechanistically modeling alcohol absorption, metabolism, and excretion across multiple compartments, the framework captures the combined dynamics of these markers and enables individualized simulations based on anthropometric inputs. Trained on data from 10 studies and validated against an independent dataset, the model demonstrated the ability to reproduce biomarker patterns and distinguish between similar drinking scenarios in simulations, supporting its use as a decision-support tool for evaluating the plausibility of reported alcohol intake.