Drug-drug interaction (DDI) is a phenomenon of alteration of the pharmacological activity of a drug(s) when another drug(s) is co-administrated in case of so-called polypharmacy. There are three types of DDIs: pharmacokinetic (PK), pharmacodynamic, and pharmaceutical. PK is the most frequent type of DDI, and often it appears as a result of inhibition or induction of drug-metabolizing enzymes (DME). In this review, we summarize in silico methods that may be applied for prediction of inhibition or induction of DMEs and describe appropriate computational methods for DDI prediction, show the current place and perspectives of these approaches in medicinal and pharmaceutical chemistry. We explained in detail sources of information on DDI that can be used in pharmaceutical investigations, medicinal practice and in computational models creation, discuss the problem of inaccuracy and redundancy of this data. In the description of methods of computer-aided prediction of DDI we provide information on the state-of-the-art in physiologically based pharmacokinetic (PBPK) computational approaches and in protein-based in silico methods dealing with DMEs. In the section of ligand-based methods we described utilization of pharmacophore models, molecular fields analysis, quantitative structure-activity relationships (QSAR), and similarity analysis applied to prediction of DDI related to inhibition or induction of DME. In conclusion, we discuss the problems of DDI’s severity assessment, mention factors that influence on the severity, highlight the issues, perspectives and practical using of in silico methods.