Using Biber's MD model, this paper investigates the register variation in crisis translation and discovers that learned exposition is the text type that comes closest to crisis translation. Crisis translation has “explicit” and “informational” features, according to statistics. When compared to the non-translational “learned and scientific” corpus, the translated corpus includes feature of “nominalization.” Furthermore, the researchers discovered that in Biber's MD model, the variables AWL and PHC had a strong liner connection with NOMZ.