This study identified factors contributing to skeletal relapse in the two-jaw surgery treatment of mandibular prognathism. A set of three standardized lateral cephalograms (T1: before surgery, T2: immediately after surgery, T3: final follow-up after surgery) were obtained from 35 patients. The surgical changes were defined as follows: postsurgical immediate change (T2–T1), postoperative stability (T3–T2) and the final surgical change (T3–T1). The occlusal plane and gonial angles were also measured. Relapse was defined as the reverse movements of the menton point (Me) and point A, with the null hypothesis stating that Me and point A do not significantly change at the postoperative stability (T3–T2). A paired t test and Pearson's correlation were used for statistical analysis. The immediate postoperative changes (T2–T1) in Me and point A were significant, and were measured to be 8.5mm backward and 3.0mm forward, respectively. Additionally, the occlusal plane and gonial angles significantly increased by 2° and decreased by 2°, respectively. The final postoperative changes (T3–T1) in Me and point A were also significant, and were measured to be 5.2mm backward and 2.5 forward, respectively; the occlusal plane and gonial angles also increased nonsignificantly by 0.6° and 0.7°, respectively. Upon investigating postoperative stability (T3–T2), Me was measured to be significantly 3.3mm forward and 1.4mm upward, whereas point A was measured to be nonsignificantly 0.5mm backward and 0.9mm upward. Therefore, the null hypothesis was rejected. Pearson's correlation showed that horizontal Me (T3–T2) and point A (T3–T2) were significantly correlated with the amounts of setback Me (T2–T1) and advancement A (T2–T1), respectively. In conclusion, skeletal relapses are significantly correlated with the amounts of mandibular setback and maxillary advancement.