Scholarly interest in legislative translation has grown substantially over recent decades, with corpus-based approaches contributing much to our understanding of the relationship between translated legislation and source texts, on the one hand, and translated and non-translated legislative texts in the target language, on the other. To date, however, most studies have been conducted on European languages. This study represents a first attempt to use corpus techniques to explore legislative translation from English into Thai. Drawing on a purpose-built, 400,000-word, parallel corpus of international treaties translated from English into Thai, and a one million-word monolingual corpus of legislative texts originally written in Thai, both of which are managed using the Sketch Engine platform, we investigate how selected linguistic features—including instances of passive voice, deontic modality and cohesive conjunction—are translated into Thai. We analyze the inter-linguistic and intra-linguistic differences we find in the light of the possible adoption of ‘plain writing’ principles (Williams 2009) in Thai legislation, the legal authenticity of the Thai non-translated texts, previously posited ‘features of translation’ including Teich’s (2003) notion of source texts ‘shining through’, Becher’s (2010) approach to explicitation and the presence of ‘unique items’ (Tirkkonen-Condit 2002), and drawing on Biel’s (2014) concept of ‘textual fit’.