Objective · To compare the baseline characteristics and treatment outcomes of mechanical thrombectomy in patients with intracranial arterial occlusion caused by embolic stroke of undetermined source and cardiogenic stroke. Methods · Retrospective analysis was made on ESUS and CS patients in registration databases who received thrombectomy in two stroke centers, Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine and Lishui Central Hospital, Zhejiang University from November 2012 to April 2019. T-test and Mann-Whitney U test were used to compare the measurement data, χ2 test and Fisher's exact test were used to compare the counting data, and the independent prognostic risk factors were analyzed by Logistic regression. Results · In all, 117 participants were eventually enrolled, including 30 (25.6%) with ESUS and 87 (74.4%) with CS. Compared with the CS group, the ESUS group was significantly younger (mean ages, 64 years vs 75 years, P=0.003) with lower median baseline NIHSS scores (12 vs. 15, P=0.020), lower median NIHSS scores at 24 h (10 vs 12, P=0.033) and lower median MRS scores at 90 days (2 vs 4, P=0.015). The rates of successful recanalization were similar. Logistic regression analysis showed hypertension (OR=0.264, 95%CI 0.099-0.704, P=0.008) and baseline NIHSS scores (OR=0.758, 95%CI 0.673-0.853, P=0.000) were independent risk factors affecting prognoses. Conclusion · Compared with CS, ESUS patients are relatively younger and have milder neurological dysfunction at onset and better prognoses; however, both groups have high mortality rates. The successful recanalization rates for mechanical thrombectomy are similar. The baseline NIHSS score and hypertension are independent prognostic risk factors.