A four-year research project started in 2012 to develop margin assessment methodologies against external hazards as well as probabilistic risk assessment (PRA) methodologies. In this project, only the decay heat removal function was taken into account assuming no loss of reactor shutdown function. The developed methodology is applied mainly for sodium-cooled fast reactors (SFRs), though it would also be applicable for light water reactors (LWRs) basically. Typical SFR heat sink is air, which is different from the heat sink in LWRs. Therefore, it is important external hazards that influence to air coolers which are located at high elevation. This project addresses extreme weathers (snow, tornado, wind and rainfall), volcanic phenomena and forest fire as representative external hazards. In this study, the external hazard evaluation, the accident sequence and the methodologies of both PRA and margin assessment are developed for each external hazard. This paper describes mainly snow margin assessment methodology development in addition to the project overview including the scope and methodologies against snow, tornado, wind, volcanic eruption and forest fire. For the snow margin assessment, the index is a combination of a snowfall speed and duration. Since snow removal can be expected during the snowfall, the developed snow margin assessment methodology is such that the margin was regarded as the snowfall duration up to the decay heat removal failure which was defined as when the snow removal rate was smaller than the snowfall speed.