In Mongolia, livestock animals were seriously killed due to effects of climate change: abnormal weather, Zud. Thus, it is required to reduce its risk by early adaptation. To limit stocking rate, pre-emptive sales triggered by insufficient rainfall, and variable sales and stocking rate regimes determined by the current season's rainfall are early adaptations. In order to implement a policy of them, it is significant to evaluate vulnerability of socio-ecological systems in the Sum (town or city) scale of Mongolia that is the finest scale of livestock animals’ management. In recent studies, there was a lack detail or specifics in economics. We examined a range of tracking policies, designed to tackle climatic variation, using a simulation model of a semi-arid grazing system, SimSAGS. In the result of simulation, it was clarified that although the flexible stocking strategies did reduce mortality losses, compared with fixed stocking, they increased average annual sales unlike with recent result of other areas’ simulation. In this study area, Zud is often observed nowadays, thus it was also clarified that was difficult to track. This result has a possibility to help their life with livestock animals, though we have to think about management of destocking for realizing these policies.