Introduction: The non-invasive assessment of body composition is a valuable, informative tool for nutritional evaluation in a wide range of settings. Bioelectrical Impedance Analysis (BIA) technology has become very popular in recent years, recommended for this purpose primarily due to its ease of use and affordability. Nevertheless, many doubts remain regarding its physiological significance and data reliability. Dual-energy X-ray absorptiometry (DXA), on the other hand, is regarded as the gold standard for body composition evaluation, although (equipment) cost and time-consuming operation can limit its use in routine clinical practice or large epidemiological surveys. Aims / Objectives: To examine and compare the total body composition in healthy university students using two different technologies – BIA and DXA. Methods: This pilot cross-sectional study involved 25 individuals, 88% women (23 years old + 6,39) and a mean BMI of 21.86 kg/m2. Body composition was assessed using a DXA Lunar Prodigy Advance (General Electric Healthcare®) and a BIA (Tanita TBF 300®). Other descriptive variables were also collected by trained dietitians, including weight, height, abdominal circumference, smoking status, sleeping hours, intestinal and urinary rhythm, and physical activity practice. Procedures respected all the principles of good clinical practice for human studies research. Statistical analysis was performed using SPSS software, and all statistical tests (descriptive and regression analysis) were two-tailed and the significance level was set at p<0.05. Results and Discussion: A weak correlation of fat mass (FM) detected with DXA and BIA (r=0,8639, P<0,01) was found. (Fig.1). In absolute terms BIA underestimated FM by 15.5%, in comparison to DXA. In contrast, a higher concordance was observed (r=0.97) for fat-free mass (FFM), however, these results were not statistically significant. Similar results were previously reported, indicating that a correlation between BIA and DXA is not always accompanied by sufficiently reliable limits of agreement between methods. Conclusion: In the present experimental conditions, there is a lack of agreement between  FM and FFM as assessed by BIA and DXA.  More work is needed to know how to correlate data from one versus the other.