||Understanding the diversity of current states, life cycles and past trajectories of households and agroecosystems is essential to contextualise the co-design of more sustainable agroecosystems. The objective of this paper was to document and analyse current states, trajectories of changes and their major drivers of households in a highly populated maize-based agroecosystem of Western Kenya. In 2013, we revisited 20 rural households that were surveyed, analysed and categorised 10 years ago (2003) in order to describe major changes in livelihood strategies, land use and soil fertility status. The household-level analysis was complemented with the analysis of secondary data on changes in drivers at the national level for the study period. The diachronic study showed a close association between drivers such as market and transport development, and the structure of rural households in terms of demographic shifts, land and labour exchanges, increased costs of agricultural inputs and better connectivity to markets. Between 2003 and 2013, the surveyed households experienced an increase in non-agricultural income by 30 %; intensity of land cultivation by 60 %; use of hybrid maize seeds by 35 %; and of synthetic fertilisers by almost 50 %. Local households increase their number of cross-bred livestock in detriment of local breeds and used less manure to fertilise their soils. In contrast, there were few changes in terms of food self-sufficiency (around 9 months per year) and in soil conditions (soil fertility was already poor in 2003). In terms of livelihood strategies, better-endowed households tended to diversify and acquire land that enabled them to adapt and benefit from the major changes observed in external drivers. In contrast, more vulnerable households sold labour and land to cope with such changes, remaining in a poverty trap. Households combine and explore diverse strategies to act, cope and adapt to fast-changing local and regional drivers. Policy or development programmes need to account for such diversity and dynamics to support the co-development of more adaptive and sustainable smallholder agroecosystems.