Due to their increasing popularity, the optimization of swarm collaborative behaviours has a high applicability potential. In the presented paper we implement a novel method of fuzzy decision-making for swarming operation, based on the concept of artificial empathy and imprecise cognition. The model is based on human learning and adaptation processes, known from neuroscience. Cognitive empathy is used, modelled as a fuzzy communication and prediction model, encompassing both imprecision and uncertainty. The main goal of the paper is to introduce a swarm example, where an agent has to decide whether to continue their own action or to collaborate with others, based on imprecise data and an unknown future. We use fuzzy similarity measures to compare the agent’s predicted future state, the empathy target’s future state and the agent’s projection of self into the future state of the empathy target. All of the predictions are assigned a reward, dependent on the goal set for the whole swarm. The decision is then fed to the lower decision module, which picks the particular actions to be performed by the platform to get into the preferred state. The paper shows a physical implementation of the model into a swarm of mobile robots. Results of improving swarm cooperation are presented.