While the identification of dynamics of land management remains a challenge (Rounsevell et al., 2012), our study contribute to improve the understanding of land-cover modification which is partly driven by changes in agricultural land management practices. The objective of this research was to investigate the subtle changes in cropland in relation to crop sequences in French mainland. All of 430 French agricultural districts were incorporated in this investigation. A powerful statistical method, Partial Triadic Analysis (PTA) was applied to assess the spatio-temporal patterns of crop sequences. The historical land-cover data (1992-2003) that we used, Teruti, consisted of 555,382 sampling points across the entire territory and used a detailed nomenclature (81 land-covers) among which 41 crops were distinguished. Crop sequences were analyzed by dividing the twelve years into three sub-periods (1992-1994, 1997-1999 and 2001-2003). The division of the whole period into three sub-periods was defined considering the potential evolution of cropping systems related to the common Agricutlural Policy reforms in 1992 and 2003. The occurences of 3-year crop sequences were then counted within each district and each sub-period. We further explored the dynamics of crop sequence patterns by applying PTA on the crop sequences datasets. PTA allowed extracting the crop sequence patterns that were common to 3 sub-periods and identifying which sub-periods differed from the common model and which crop sequences were responsible for these discrepancies. The result of PTA on both permanent and annual crops included crop sequence datasets showed that crop sequence patterns were roughly structured in vineyard-based, maize monoculture-based and annual crops-based cropping systems. The second PTA on annual crops related crop sequences (maize monoculture not included) datasets revealed cropping systems diversity across the territory. The occurence of stability and dynamics of crop sequence patterns during the study period was greatly contrasted between cropping systems e.g. more temporal dynamics revealed in the districts where cropping system was based on oilseed crops (sunflower and rapeseed) and wheat than the districts where the cropping system was based on maize. It revealed a common temporal dynamics of crop sequence patterns in districts belonging to the same cropping system e.g. the increase in rapeseed-wheat based crop sequences areas during the first two sub-periods could be observed in most of rapeseed cultivation zone. Our approach may have implications for assessing production system performance by monitoring the cropland change in other countries. It could be also applied using time series remote sensed data.