In the present paper we have faced the problem of estimating the mass distribution of groups of galaxies, which is of crucial interest for galaxy evolution and cosmological models. Recent observational works and surveys have made available several catalogs of groups obtained by using various and rather different algorithms. Moreover, the mass of groups can be estimated in several ways. The aim of our work is to analyze to what extent the observational estimate of the group mass distribution is sensitive to the features of the catalogs and, in particular, to the mass estimators. We believe that this is a crucial check in order to set valuable constraints on the theoretical model parameters. Our main results indicate that several estimators that we have considered give nearly the same mass distribution. On the contrary, there is a significant dependence of the distribution of the mass and other dynamical parameters of groups on the identification algorithm adopted to compile the catalogs. Finally, we take into account this algorithm dependence in the comparison between the observational mass distribution and the Press & Schechter and Peebles theoretical distributions.