This study was planned to investigate the parameters detecting risk of developing atrial fibrillation (AF) in patients with sinus rhythm with structural heart disease. Forty-five patients with AF and 37 patients without AF but with structural heart disease (Group I) were included in this study. Thirty-eight patients (Group II) had successfully undergone medically or electrically cardioversion after transesophageal echocardiography. The restoration of sinus rhythm could not be achieved in 7 patients who were excluded from this study. After providing sinus rhythm, amiodarone was given orally to the patients to prevent recurrences. Left ventricular ejection fraction (LVEF) was calculated and left atrial diameter (LAD) was measured by echocardiography in group I and in group II after cardioversion. A 12-lead electrocardiography (ECG) was simultaneously obtained from all the patients. In these ECG recordings, maximum P wave duration (P max), minimum P wave duration (P min), and P wave dispersion (P dispersion) were calculated. P dispersion was expressed as "P max-P min." Also, the highest P wave voltage is expressed as P amplitude maximum (P amp max), the lowest P wave as P amplitude minimum (P amp min), and P amplitude dispersion (P amp dispersion) was calculated as the difference of both. In univariate analysis, P max, P dispersion, P amp max, P amp dispersion, LAD, LVEF, and old age were significant predictors of chronic AF (p < 0.001, p < 0.01, p < 0.01, p < 0.01, p = 0.003, p = 0.02, and p = 0.01, respectively). However, in multivariate analysis, P max and LAD were independent predictors of chronic AF in patients with structural heart disease (r = 0.39, p < 0.05; r = 0.34; p < 0.05, respectively). In conclusion, in estimating the risk of developing chronic AF, P max and LAD are predictive parameters in patients with sinus rhythm with structural heart disease.