P-glycoprotein (P-gp) is an ATP dependent efflux transporter protein that has been demonstrated to play a critical role in affecting the absorption, metabolism, elimination and toxicity (ADMET) characteristics of a large number of marketed drugs. Therefore, it is important to evaluate whether or not compounds of interest are likely to interact with P-gp and/or other efflux transporters. An in silico efflux substrate (potential substrate of P-gp and or other transporters) classification model has been developed based on in vitro bi-directional Caco-2 cell permeability and five descriptors, using 14 marketed drugs and >100 discovery compounds synthesized at Bristol-Myers Squibb PRI. The model suggests that efflux substrates tend to contain electron deficient aromatic rings, are highly branched, and most contain tertiary nitrogen. This model demonstrated approximately 80% predictability of both non-substrates and substrates from a training set of 125 compounds. For a validation set of 46 compounds the predictability was approximately 72% for non-substrates and approximately 89% for substrates. The model has the potential to be used both as a filter for library designs to identify potential efflux substrates in early discovery as well as a primary screening methodology to identify the efflux substrate potential of drug candidates.