This thesis was developed in the Scientific, Mathematical and Technological Education line of the Graduate Program in Education, of the Faculty of Education of the University of São Paulo. It is a research that aims to investigate the views and movements of leaders of research groups in Ethnomathematics in Brazil, in the initial training of teachers who teach mathematics. In the theoretical framework, it presents Paulo Freire's studies in the discussions about the formation of teachers. Regarding the Ethnomathematics Program, the theorizations proposed by Ubiratan D'Ambrosio and the considerations of Maria do Carmo Santos Domite on the formation of teachers in the perspective of Ethnomathematics are brought up. This research has a qualitative methodological approach, and the data were collected through semistructured interviews, with questions elaborated from a Domite"s study, contextualizing the research question. The interviewees are the leaderscoordinators or vice-coordinators -of the eight Brazilian groups registered in the CNPq group directory, which includes the word "Ethnomathematics" in its name. These groups make up all five regions of Brazil. The NVIVO12 software was used to assist in the identification of the sets (categories), through these words, the most frequent in the speech of the interviewees: Mathematics, Ethnomathematics, Knowledge and Teacher. Such words were obtained through the word cloud and considering our subjectivity as researchers, we arrived at these analysis categories. "Different viewpoints on Mathematics", "First knowledge and scientific knowledge" and "Ethnomathematics in the initial training of teachers who teach Mathematics". From this process, these subcategories emerged "Mathematics as social and cultural production", "interchange between school knowledge and everyday knowledge" and "Ethnomathematics Stance". The work considers the importance of the teacher trainer through the dialogue of the knowledge produced socially and culturally in the daily practices of the future teachers with the academic knowledge, without there being imposition of some knowledge about the other, assuming an ethnomathematics posture in their practice and always respecting and valuing different branches of mathematics.