Demographic literature is rich of empirical analyses showing how women have historically experienced lower mortality rates than men. In this paper, we consider a measure of the gender gap in mortality rates, the Gender Gap Ratio, across a wide range of populations collected in the Human Mortality Database. With the aim of highlighting similarities and differences between the countries considered, we apply a functional clustering method to the multivariate time series of Gender Gap. We reconstruct the functional form of the trends from the available discrete observations and derive the curves through non-parametric smoothing. Results for 65-years-old people from 1965 to 2014 are presented and discussed.