At the building scale, the use of green roof has shown a positive impact on urban runoff (decrease and slowdown in peak discharge, decrease in runoff volume). The present work aims to study whether similar effects are possible at the basin scale and what is the minimum spreading of green runoff needed to observe significant impacts. It is particularly focused on the circumstances of such impacts and how they can contribute to storm water management in urban environment. Based on observations on experimental green roofs, a conceptual model has been developed and integrated into the SWMM urban rainfall-runoff model to reproduce the hydrological behaviour of two different types of green roof. It has been combined with a method defining green roofing scenarios by estimating the maximum roof area that can be covered. This methodology has been applied on a long time series (18 years) to the Châtillon urban basin (Haut-de-Seine county, France) frequently affected by urban flooding. For comparison, the same methodology has been applied at the building scale and a complementary analysis has been conducted to study which hydrometeorological variables may affect the magnitude of these hydrological impacts at both scales. The results show green roofs, when they are widely implemented, can affect urban runoff in terms of peak discharge and volume, and avoid flooding in several cases. Both precipitation – generally accumulated during the whole event-and the initial substrate saturation are likely to have an impact on green roof effects. In this context, the studied green roofs seem useful to mitigate the effects of usual rainfall events but turn out being less helpful for the more severe ones. We conclude that, combined with other infrastructures, green roofs represent an interesting contribution to urban water management in the future.