Neuromagic is underpinned by the union of neuroscience and magic. However, the contribution between them is unbalanced. While neuroscience has used magic tricks to study different cognitive processes and develop sophisticated experiments, the returned contribution has not penetrated into magicians praxis. In neuroscience, a priori theories are supported by objective experimental data. However, this steps are switched in magic: subjective experimental evidence generates a posteriori theories. This scientifically invalid methodology leads, as a consequence, to several unvalidated theories about the same topic. To solve this problem, testing this theories with previously validated brain models is essential, as it will separate the ones that can be reproduced with realistic models from the ones that can not. Here, we use a popular brain model -the bump attractor- to describe the most studied magical principle: misdirection. Misdirection aims to avoid the detection of tricky maneuvers by affecting the attention and the memory of the spectators. Misdirection is been defined and classified in multiple ways for different magicians. For the first time, we will support with a model of neural activity, a recent misdirection classification (Bestue, 2019). The model reproduced the efficiency and provided a neural explanation for each of the three different subcategories of misdirection proposed by the classification: “direction”, “division” and “diminishment”. That way, we provided first neuroscientific evidence for a misdirection classification. We hope this work motivates both magicians and neuroscientists to provide detailed models to different magical principles, so they rely in objective brain functioning and not in subjective psychological theories.