This paper seeks to identify the determinants of the choice of treatment of pregnant women in Cameroon. Theoretically, the methodology is based on a discrete choice model with random utility. Empirically, the econometric specification is a Nested Multinomial Logit Model. The data used comes from the Demographic Health Survey (DHS) organized in 2011 by the National Institute of Statistics. The results reveal that uneducated women or those having only a primary education prefer to meet the traditional midwives than seek modern maternal health services. Moreover, the absence of a paid job for the pregnant woman, the large size of the household, Islamic or animist religion, poverty, high costs of healthcare and transportation are constraints which make the pregnant woman to prefer the services of traditional midwives to modern services of maternal health. The use of modern healthcare services by pregnant women in Cameroon can therefore be improved by at least two means: firstly, by improving on the level of education of women and economically empowering them. Secondly, in a context where the costs of healthcare services are paid directly by the pregnant women themselves or by their families, it is important to put in place health insurance schemes in order to guarantee proper follow-up of pregnant women until delivery as well as taking care of complicated cases.