In this work, a two-step sequential extraction scheme for the determination of trace elements in Arctic PM10 samples was optimized by using two certified reference materials (CRMs). By means of an experimental design for qualitative variables, the five most common extracting solutions for particulate matter (PM) sequential extraction (high purity water (HPW), 0.032 M HNO3, 0.022 M HCl, 0.11 M CH3COOH, and 0.012 M CH3COOH/CH3COONH4 buffer) and two different extraction methods (stirring and ultrasounds) were compared. The purpose of the study was the identification of the procedure which gives the best estimation of the anthropogenic portion of the elements present in PM10 samples. The use of ultrasounds instead of stirring induced a low but significant decrease of the extraction of all the elements and a decrease in the repeatability of the procedure. Diluted HNO3 was the extractant which allowed to maximize the extraction of anthropogenic elements (As, Cd, Pb, Zn) with respect to crustal ones (Al, Si, Ti). The optimized procedure proved successful in avoiding contaminations and, therefore, suitable to be applied to PM samples having extremely low concentrations, such as samples collected in polar or other remote areas. The chosen procedure was applied to ten Arctic PM10 samples, allowing for a better identification of their sources. Indeed, it was possible to hypothesize that even though the concentrations of As, Cd, K, Mg, Mn, and Ni in spring and summer were different, their mobility and, therefore, their chemical form in the analyzed PM10 samples were probably similar.