The aim of this work is to assess the potential of spatial databases to monitor soil quality at a national level. Data were collected in the framework of French NationalProgrammes. To address this issue we show that the spatio-temporal statistical analyses must be adapted to the sampling design and to the nature of the information studied. Among the set of the possible soil variables, this work focused on some of those having a high environmental impact: the organic carbon contentand the content of some trace elements. Our results show the possibility to map soil properties at national scale, to reveal strong spatial structures and, to attribute them to different natural and artificial processes. Large temporal trends could also be detected and explained. I discuss the limitation of the present designs and of the statistical analyses we conducted and I propose fudher research developments for monitoring of soil quality.