Digital Twins (DT) able of reproducing patient-specific electrical patterns during atrial fibrillation (AF) could help identify optimal therapy. 12 anatomically and electrophysiologically customized DTs were constructed. From CT images the left atrium was reconstructed to which anatomical features were added. The region-specific cycle length (CL) and conduction velocity (CV) were estimated from intracavitary recordings. These biomarkers were translated into parameters that could be used in simulations (diffusion and electrical remodeling) from a calibration study based on simulations in an atrial atlas. To compare the results, 12 basic Digital Twins that reproduce only the patient's anatomy were used. The DTs faithfully reproduced, in the simulations carried out, the biomarkers measured in patients, not only globally but also regionally: 194 ± 15 vs 192 ± 10 ms for CL, 376 ± 146 vs 372 ± 54 mm/s for CV, for an example case. Significant differences were found in the relative errors obtained with the DTs versus the basicDTs both in terms of CL (0.14 ± 0.07 vs. 0.23 ± 0.09, p < 0.05) and CV (0.37 ± 0.12 vs. 0.54 ± 0.15, p < 0.05).