Trace element analysis of human hair is widely used to provide an indicator of body burden, but there is a major problem associated with preparation. The washing procedure adopted before analysis represents still a critical point which deserves particular attention. This study aimed at comparing the efficiency of the most commonly used washing methods to identify the procedure which allows for satisfactory removal of external contamination alone. As results depend strongly on the element, toxic (Cd and Pb), essential trace (Cr, Cu, Se and Zn) and major (Ca, K and Mg) elements were tested. The comparison was carried out with six different methods which include use of ionic and non-ionic detergents, organic solvents, chelating agents and sonication. Removal efficiency varied according to the element and pretreatment, the highest being observed for EDTA and the lowest for sonication. Unsatisfactory results were found using an acetone/methanol mixture for the potential contamination caused by it. Organic solvents showed the highest removal efficiency for K and toxic elements, and seemed to be effective in removing a limited but significant fraction of element incorporated in the lipid matter of hair. The ionic detergent Na lauryl sulphate, instead, was more effective for essential elements, but also triggered a higher analytic variability. As detergents appear to remove only external contamination, a non-ionic detergent such as Triton X-100 is proposed as a reliable alternative to the acetone method recommended by IAEA. Practical advantages are safe working conditions, rapidity of application and reduction in costs.