Protein precipitation is the most commom sample preparation method applied in studies involving biological matrices, such as plasma and serum, in metabolomics. However, a poor protein precipitation step may reduce the lifetime of chromatographic columns, may cause interferences in Mass Spectrometry, and may result in limited access to sample metabolites. In this work different protein precipitation approaches were evaluated in order to assess the best protocol for untargeted metabolomics studies. For this purpose, solvents selection; sample volumes; sample/solvent ratio and temperature were evaluated by a factorial planning. Serum samples were analyzed by Bradford assay.