There are many alternatives to soya for animal feed in the tropics (e.g. meals coming from local protein-rich resources such as beans, peas, aquatic plants or leguminous foliage). The study and evaluation of such alternatives must be based on their amino acid availability and profile, using the approach of investigating non-ruminant species. For this criterion, there is often a gap between the amino acid profiles of plant resources and the profile of amino acids truly available for the animal. In the short and medium term, new studies have to be performed to take into account the large diversity of rich protein resources in the tropics. Overall, the alternatives are broader for herbivores than for other animal species, since the concentration of fibre and secondary compounds is a limiting factor that will discriminate their proposal of use among domestic animal species. Such evaluations must also take into account the farming system functions and productive purposes. The paradigm is changing and, compared to years ago, maximising animal performance is not the priority goal of the systemic approach. Currently, multiple animal responses to alternative diets should be taken into account for their optimisation. A criterion such as environmental impact is often decisive in the combination of global and local approaches.