At recent regular sessions of IMO (International Maritime Organization) /MSC (Maritime Safety Committee) several safety regulations have been discussed by FSA (Formal Safety Assessment) procedures. In the discussion the problems based on statistical uncertainty were arisen. Since member governments of IMO have a tendency to treat the problems arbitrarily, there exist differences among states of the IMO member governments In this paper scientific treatment methods of problems based on statistical uncertainties are presented and applied to two actual problems which were arisen during the process of discussion in the continuous session of IMO/MSC. The one is the treatment of unknown caused accidents, the other is related to validity of the indication based on risk reduction by introduction of new regulation. The former is resolved by application of Bayesian statistics, the latter should be treated by reliability analysis. The methods presented in this paper will be the tool for formation of mutual agreement among stakeholders who have opposite interests.