Natural disaster risk is the possibility of disaster-related losses depending on hazard, vulnerability, exposure and emergency response and recovery capability. Vulnerability is the indicator measuring the damage extent of hazard-affected body. Vulnerability is not only an important part of risk assessment for loss estimation and disaster, but also the link between hazard-inducing factors and disaster. This study collected meteorological, soil, land use and field management data along with other related information on the research area to evaluate the vulnerability of maize drought disaster based on CERES-Maize model in western Jilin Province. According to the define of vulnerability of natural disaster risk and climate change, the study established the evaluation method of drought vulnerability based on disturbance, sensitivity and adaptive capacity. Data for three typical drought years (2004, 2006 and 2007) in western Jilin Province were used to calculate drought vulnerability using CERES-Maize model. Regression analysis was also conducted for maize yield losses caused by drought and vulnerability index in the three typical drought years for each growth period. The results showed exponential correlations between yield loss and vulnerability, which were significant at alpha = 0.05 (F test) for different growth stages. This indicated that it was reasonable to evaluate and predict maize vulnerability to drought using the established model in the region. Correlation coefficients indicated most close relationship between maize yield losses and vulnerability index at tasseling to milk-ripe stage and jointing to tasseling stage, and followed by milk-ripe to maturity stage and seeding to jointing stage. Drought vulnerability indexes of maize were divided into 4 grades and draught vulnerability zone maps of western Jilin Province drawn on GIS platform. The results showed that areas with high drought vulnerability included Baicheng, Taonan and Zhenlai. Low drought vulnerability areas included Songyuan and Fuyu. The established drought vulnerability evaluation model was suitable for evaluating and predicting drought vulnerability of maize at different growth stages and maize yield loss due to drought. The results of this study provided the basis for improving agricultural drought risk and emergency response and recovery capability.