Objective. This study aims to evaluate the combination of genioglossus (GG) activity and anatomical characteristics in predicting outcomes of velopharyngeal surgery in patients with obstructive sleep apnea (OSA). Study Design. Case series with planned data collection. Setting. Sleep medical center. Subjects and Methods. Forty patients with OSA underwent overnight polysomnography with synchronous genioglossus electromyography (GGEMG) using intraoral electrodes. The upper airway anatomy was evaluated by 3-dimensional computed tomography in patients with OSA. All patients received the same type of velopharyngeal surgery, consisting of revised uvulopalatopharyngoplasty with uvula preservation and concurrent transpalatal advancement pharyngoplasty. We followed up all patients using polysomnography for at least 3 months postoperatively. Results. Twenty-five patients (62.50%) were responders, and 15 patients (37.50%) were nonresponders. The decreased apneahypopnea index was significantly positively correlated to the sleep onset GGEMG (P =.006) but was negatively correlated to the change in GGEMG (P =.013) and tonic GGEMG (P =.018). Multiple regression analysis revealed that the minimal crosssectional airway area at the velopharynx (VmCSA) (odds ratio [OR], 1.760; P =.019) and the sleep onset GGEMG (OR, 0.322; P =.043) were significant predictors for surgical outcomes. Combined the two predictors, the area under the ROC curve was 0.901 (OR, 0.789; P =.001) for surgical success, was more valuable than any one predictor. The area under the ROC curve with GGEMG was 0.843, VmCSA was 0.848. Conclusions. The combination of sleep onset GGEMG and VmCSA can predict the outcome of velopharyngeal surgery in patients with OSA.