The occurrence of atrial fibrillation (AF), one of the most socially significant arrhythmias, is associated with the presence of areas of fibrosis. Fibrosis introduces conduction heterogeneity into the cardiac tissue and, thus, may be a substrate for spiral wave reentry, which provokes the onset of AF and is often associated with its persistence. Despite results from computer and animal models of cardiac tissues, data on the conditions under which microreentries occur in human tissues are limited. In this work, we conducted a study of the new approach to modeling the fibrous atrial tissue, which takes into account the cellular structure and conduction in fibrosis areas. Using the Potts model, we created a realistic texture of atrial tissues remodeled by fibroblasts and showed the presence of pathways in such a system with a low proportion of fibroblasts. Our study revealed the relationship between the shape of the cells’ action potential, their location in the tissue, and the direction of the wave propagation. The wavefront obtained in the model creates a dynamic heterogeneity of the tissue, which affects the migration and pinning of spiral waves, and explains the formation of microreentries in the cardiac tissue. In the future, such a model can become a potential tool for predictive modeling of AF and the search for ablation target identification.