Journal of Intelligent Computing Fillipe Madureira, Hendrik Macedo 8 4 2017 http://www.dline.info/jic/fulltext/v8n4/jicv8n4_3.pdf Machine Translation is a highly demanding research field in natural language processing. The goal is to investigate a model which can automatically translate a document from one language to another. The mainstream approach for that is the Phrase-based Machine Translation (PBMT). In [9], a innovative method based on distributed representations was proposed to automate the process of generating dictionaries and phrase tables of PBMT-based solutions, that primarily rely on raw counts. In this paper, as in the Mikolov’s, we have used distributed representations of words to perform automatic machine translation between languages but, unlike it, we have included the Portuguese language in the analysis which, actually, is the primary goal of our work. We have properly evaluated the performance in regards to Precision metrics. Results shows a precision of up to 89% if we consider translation between the Portuguese and the English.