The decoding of the DNA structure and development of new molecular methods of its analysis, as well as identification of specific genomic changes responsible for malignant transformation, have become the turning points in elaboration of novel anti-tumour drugs directed against molecular and genetic targets of tumor growth. Transition from empirical screening of agents inhibiting tumour cell proliferation to molecule-targeted analytical methods has raised a number of serious methodological issues regarding preclinical evaluation of novel medicines. The objective of this paper was to analyse general principles and features of preclinical efficacy and safety studies of different classes of modern anti-tumour drugs with a view to improve existing national guidelines. The paper reviews various aspects of preclinical studies of different classes of anti-tumour drugs (small molecule chemotherapy drugs, hormones and hormone antagonists, alkylating agents and antimetabolites, microbial and herbal medicines, as well as monoclonal antibodies). The article explores general principles of studying the drugs’ pharmacological activity in vitro, ex vivo, and in vivo, and evaluating their pharmacokinetic parameters. It describes various methods and models of research, summarises specific aspects of determination of genotoxicity, carcinogenicity, reproductive toxicity, mutagenicity, acute and chronic toxicity of various groups of medicines. It also lists criteria for selecting drug doses for toxicokinetic studies. The need for harmonisation of national requirements for conducting preclinical studies with the European standards entails alignment of terminology and further development of general algorithms for selecting doses and determining the necessary scope of research. The use of biomarkers in preclinical studies will make it possible to exclude inefficient compounds from further research.