The electron beam treatment planning systems, with the availability of more and more developed hardware for complex calculation algorithms, need quality controls of accuracy and comparability of treatment plans from different systems. In this paper three systems have been evaluated; their similar computation algorithms (pencil beam) have been used but in some conditions (beam reconstruction, oblique incidence and dishomogeneity) different results have been drawn. The basic data (PDD, dose profile, etc.) were the same for every system, therefore it has been observed that different results are mainly due to differences in: a) input of parameter values, b) calculation algorithm approximations, c) image acquisitions and elaboration.