Abstract We have developed a new simulation technique to model zooplankton behaviour. The approach utilizes neither the conventional artificial intelligence nor neural network methods. We have designed an adaptive behaviour network, which is similar to Beer [(1990) Intelligence as an adaptive behaviour: an experiment in computational neuroethology, Academic Press], based on observational studies of zooplankton behaviour. The proposed method is compared with non- “intelligent” models—random walk and correlated walk models—as well as observed behaviour in a laboratory tank. Although the network is simple, the model exhibits rich behavioural patterns similar to live copepods.