Pulmonary embolism (PE) is associated with atrial fibrillation (AF). This study sought to explore if the CHADS2 score could predict the prognosis of PE in patients with AF. In a tertiary hospital, 4,288 consecutive patients with diagnosis of PE were screened. In total, 305 patients with PE had AF and were included in this retrospective study. In-hospital outcome was defined as at least one of the following: death from any cause, need for intravenous catecholamine administration, endotracheal intubation, cardiopulmonary resuscitation, or thrombolytic therapy. The in-hospital outcome occurred in 10.2% of the patients. Patients with adverse outcome had higher CHADS2 score, CHA2DS2-VASc score, and simplified pulmonary embolism severity index (sPESI) score. The area under the receiver operating characteristics curve was 0.66, 0.62, and 0.71 for CHADS2 score, CHA2DS2-VASc score, and sPESI score, respectively, in predicting in-hospital outcome. The incidence of in-hospital outcome was 3.4 and 14.4% in sPESI = 0 and sPESI ≥1 groups (p < 0.01). CHADS2 also had good predictive value with the incidence of in-hospital outcome, being 4.6% in CHADS2 < 2 and 14.3% in CHADS2 ≥ 2 groups (p < 0.01). The incidences of in-hospital outcome were 2.6, 4.8, 7.4, and 17.3% in patients with sPESI = 0 and CHADS2 < 2, sPESI = 0 and CHADS2 ≥ 2, sPESI ≥ 1 and CHADS2 < 2, and sPESI ≥ 1 and CHADS2 ≥ 2 (p < 0.01), respectively. In multivariable analysis, CHADS2 (odds ratio: 1.50; 95% confidence interval: 1.11–2.02; p < 0.01) was an independent predictor of in-hospital adverse outcome. High CHADS2 score could predict worse in-hospital outcome in patients with PE and AF.