Due to the lack of efficacy, undesirable pharmacokinetics, and toxicity, many potential drug candidates ultimately fail to reach the market. Approximately 50% of drug research and development failures are attributed to ADMET (Absorption, Distribution, Metabolism, Excretion and Toxicity) properties. Therefore, ADMET analysis has become instrumental in drug discovery to reduce risks of failure. Given the cost and labor-intensive nature of in vivo and in vitro drug assessments, in silico techniques for predicting ADMET and other drug properties have advanced in the preclinical phase. This enables early assessments and provides initial insights for subsequent in vitro ADMET evaluations. In this chapter, the roles of ADMET in drug discovery, particularly ADMET prediction in the virtual screening process, were focused. This section introduces the bases of ADMET and the common ADMET predictive software. Additionally, it explores the applications of ADMET in virtual screening for metabolic diseases. The section also presents the challenges of ADMET modeling, accompanied by suggested solutions. Furthermore, it outlines future directions and emerging research prospects in ADMET modeling prediction for drug discovery.