In developed countries, breast cancer mortality has decreased during the last decades due to, at least partially, the advent of mammography screening. Organised programmes aim, among other objectives, to increase participation and decrease social inequalities in screening access. We aimed to characterise the evolution of socioeconomic disparities in mammography screening before and after the implementation of an organised programme in Geneva, Switzerland. We included 5345 women, aged 50-74years, without past history of breast cancer who participated in the cross-sectional Bus Santé study, between 1992 and 2014. Outcome measures were: 1) never had a mammography (1992-2014) and 2) never had a mammography or not screened in the two years before being surveyed (subgroup analysis, 2007-2014). Educational attainment was divided in three groups (primary, secondary and tertiary) and period in two (before/after introduction of a screening programme in 1999). We calculated measures of relative and absolute change, including the relative (RII) and slope (SII) indices of social inequality adjusted for age and nationality. We compared the prevalence of screening before and after screening programme implementation using Poisson models. The proportion of unscreened women decreased during the study period from 30.5% to 3.6%. Lower educated women were more frequently unscreened (RII=2.39, p<0.001; SII=0.10, p<0.001). Organised screening decreased the proportion of unscreened women independently of education (prevalence ratio _{before vs. after} =4.41, p<0.001), but absolute and relative inequalities persisted (RII=2.11, p=0.01; SII=0.04, p=0.01). Introduction of an organised programme increased women's adherence to mammography screening but did not eliminate social disparities in screening participation.