Agricultural intensification is often recognized as a major driver of the decline of wild biodiversity in farmland. However, few studies have managed to collect relevant data to link the temporal dynamics of farmland biodiversity to the characteristics of intensive agriculture over large geographical areas. We used 7 years of data from a French citizen science programme, wherein 1,216 farmers monitored biodiversity in 2,382 fields encompassing field crops, meadows, vineyards or orchards, to examine the temporal trends in abundance of five taxonomic groups of invertebrates (solitary bees, earthworms, butterflies, beetles, molluscs) and their links with agronomic practices and surrounding landscape. We observed significant temporal trends in abundance for many taxonomic groups and in many crop types. Flying taxa (solitary bees and butterflies) were generally declining, while the trends of soil taxa were more variable. Most trends were significantly related to farming practices or landscape features. We observed a negative link between use of synthetic inputs (pesticides, mineral fertilization) and the trend in abundance of flying taxa in field crops, while in meadows organic or mineral fertilization was the main explanatory practice, with contrasting relationships across taxonomic groups. Besides, the trend in abundance of beetles and molluscs was more positive in permanent versus temporary meadows. Finally, in vineyards, the trend in abundance of solitary bees was positively related to the presence of woodland in the landscape, whereas the reverse was true in meadows. Synthesis and applications. Our results provide further support for the role of citizen science as a promising source of large‐scale spatial and temporal data in farmland, contributing to the identification of agronomic practices that can help mitigate biodiversity decline. Our analyses suggest that reducing chemical inputs may not only reduce the decline in bees and butterflies, but sometimes even promote their regrowth. Increasing organic fertilization may foster bee and beetle abundance in meadows but reduce mollusc abundance, while preventing ploughing of meadows may promote soil invertebrate abundance. Finally, such citizen science programmes engage farmers to undertake monitoring. Whether such group engagement may also contribute to biodiversity conservation by raising farmers’ awareness remains to be addressed.