Oxidative degradation of lipids, especially that induced by reactive oxygen species (ROS), leads to quality deterioration of foods and cosmetics and could have harmful effects on health. Currently, a very promising way to overcome this is to use vegetable antioxidants for nutritional, therapeutic or food quality preservation purposes. A major challenge is to develop tools to assess the antioxidant capacity and real efficacy of these molecules. Many rapid in vitro tests are now available, but they are often performed in dissimilar conditions and different properties are thus frequently measured. The so-called 'direct' methods, which use oxidizable substrates, seem to be the only ones capable of measuring real antioxidant power. Some oxidizable substrates correspond to molecules or natural extracts exhibiting biological activity, such as lipids, proteins or nucleic acids, while others are model substrates that are not encountered in biological systems or foods. Only lipid oxidation and direct methods using lipid-like substrates will be discussed in this review. The main mechanisms of autoxidation and antioxidation are recapitulated, then the four components of a standard test (oxidizable substrate, medium, oxidation conditions and antioxidant) applied to a single antioxidant or complex mixtures are dealt with successively. The study is focused particularly on model lipids, but also on dietary and biological lipids isolated from their natural environment, including lipoproteins and phospholipidic membranes. Then the advantages and drawbacks of existing methods and new approaches are compared according to the context. Finally, recent trends based on the chemometric strategy are introduced as a highly promising prospect for harmonizing in vitro methods. © 2007 Elsevier Ltd All rights reserved.