The BioGeographical Ancestry (BGA) of an individual can be inferred from their Deoxyribonucleic Acid (DNA), particularly by using Single-Nucleotide Polymorphism (SNP) markers. This short paper aimed to predict the continental BGA by adopting a supervised Machine Learning (ML) method and relying on an innovative SNP panel. Starting from individuals with known BGA, a model pipeline was applied within a nested cross-validation strategy to perform model selection and assessment. The results showed a good discrimination capacity of the novel panel and plausible misclassification patterns that may be connected more to the complexity of the phenomenon rather than to inference problems, which require a discussion of the BGA uncertainty. These findings laid the groundwork for further research with the ultimate purpose of inferring BGA at a finer level.