A model of behavior of a community of relatively weak peaceful agents in the presence of a strong predatory agent is constructed and studied. The predatory agent can attack peaceful agents, kill and eat them. The internal control system of the peaceful agent is a neural network. The outputs of the neural network determine the actions of the peaceful agent. The behavior of the predatory agent is determined by simple logical rules. The model was analyzed by means of computer simulation. It is shown that with a sufficiently large amount of food for peaceful agents, the collective of peaceful agents defeats the predatory agent, namely, over time, the resource of peaceful agents steadily grows, and the resource of the predatory agent decreases to zero, i.e. the predatory agent dies. However, with a small amount of food for peaceful agents, the predator has a significant advantage over peaceful agents.