Abstract Study experience of ecologist plays an important role in assessing the contribution of different influencing factors to ecological vulnerability, helping policy makers to target measures for ecological restoration. However, uncertainty is unavoidable due to variation of study experience among experts. In this study, a new method that combines Delphi survey, geographic information system and Monte Carlo simulation was proposed to assess regional ecological vulnerability and to quantify the uncertainty of assessing result. We illustrated the capacity of this method by using a case study in northeastern Inner Mongolia, China. An index system for 13 spatial variables was established to calculate an ecological vulnerability index (EVI) from the three aspects of ecological sensitivity (ES), ecological resilience (ER) and natural-social pressure (NSP). The assessment shows that the southwestern region of the study area, especially in the counties of Sonid Left and Right, was seriously threatened by a high ES and a low ER. Onguiud county in the Greater Hinggan Mountains had a high EVI due to an intensive NSP. Based on the assessing result and regional road distribution, an EVI cost curve was created to facilitate the prioritization of allocating limited funds among the various counties for roadside ecological restoration.