The increased prevalence of overweight and obesity has become a major factor in public spending in countries around the world. The diagnosis of overweight and obesity is based on body mass index (BMI) and body fat percentage (BFP). The World Health Organization proposed BMI cut-off points to define overweight and obesity. Recently epidemiological studies established as normal BFP a BFP < 25 for men and BFP < 30 for women. A high correlation between a high BMI, abnormal BFP and skin thinness have been found in numerous studies. The aim of this work is to evaluate the k-means clustering algorithm using anthropometric measurements for the classification of subjects with overweight/obesity and abnormal BFP. Precision (P), accuracy (Acc) and recall (R) were calculated to evaluate the efficiency of the method to classify overweight/obesity and abnormal BFP. Results of this research suggest that the k-means method applied to anthropometric measurements can make an acceptable classification of overweight/obesity and abnormal BFP. The arm circumferences values show the best Acc, P and R (0.79, 0.84 and 0.71) compared to all other measurements for overweight/obesity diagnosis, otherwise, suprailiac and abdominal skinfolds values show the best Acc, P and R (0.73, 0.73 and 0.64) compared to all other measurements for abnormal BFP diagnosis. Results that are supported by studies asserting a strong relationship between arm circumferences, abdominal skinfold, suprailiac skinfold, BFP and BMI. Other machine learning techniques, such as neural networks and the support vector machine, will be studied in the future to assess the relationship between BMI, BFP and anthropometric measurements.