Aim. The aim of this study was to explore the prediction factors for incidence of acute mountain sickness (AMS) in young males newly entering highland areas. Methods. A retrospective study of 4367 records of male highland soldiers from 2000 to 2005 was done. The factors were tested by logistic regression. Results. After selection by univariate model, ethnicity, altitude, season, deployment type, and prophylaxis were inserted into a multivariate model. The adjusted odds ratio (AOR) was 0.078 for Tibetan compared with Han. AORs for altitudes 3600 to 3700, 4000 to 4300, and 4600 to 4700 m versus 2900 to 3100 m were 4.490, 4.532, and 4.964, respectively. AOR for cold season versus warm season was 1.332. AORs for emergency land deployment and air deployment versus normal land deployment were 2.261 and 1.614, respectively. The AOR was 0.741 for prophylaxis versus none. The area under receiver operating characteristic curve was 0.731 (optimal cutoff = 0.370). Conclusions. Adjusting for altitude, risk factors that contributed to AMS were being non-Tibetan, cold season, greater speed of transport, emergency conditions, and without prophylaxis. The model established is acceptable for assisting AMS prediction.