Rotors and ectopic foci are targets for ablation to revert atrial fibrillation (AF). Spectral analysis is not accurate for rotor localization during chronic AF. Studies suggest that complex fractionated atrial electrograms (CFAE) are generated by rotor tips. The aim of this work is to locate rotor tips applying approximate entropy (ApEn), which is a non-linear dynamic measure. For this, a chronic AF episode was simulated in a 3D model of human atria. Electrograms were simulated in the atrial surface and ApEn was calculated. Our results show that high ApEn values matched with rotor tips. This findings suggest that targeting ApEn areas may potentially be effective in identifying rotor tips for ablation in chronic AF.