This dissertation presents an analysis of farmer-led irrigation (FLI) along sand rivers in dryland areas of Kenya and Zimbabwe. Both study areas are characterised by a relative abundance of water and land resources, but contrast in socio-economic context. This study evaluates the operations, drivers, challenges and coping mechanisms of smallholder irrigators over time. Through surveys, in-depth interviews and mapping of farm plots, this study reveals how irrigation dynamics manifest in terms of time and space. In Zimbabwe, rural households mostly engage in irrigated farming from sand rivers as a result of multiple crises (climatic, political and economic). In Kenya, primarily market-oriented farmers venture in dynamic partnerships with capital providers on leased lands. They often change farming partnerships and plots, both out of necessity and opportunity. A longitudinal study shows how farmers frequently pause, stop or alter irrigation strategies by choice or force, confirming flexibility and autonomy as distinctive features of FLI. These findings imply that emerging irrigation policies aiming to catalyse FLI in sub-Saharan Africa need to appraise diverse farmer livelihoods. Policies that adopt FLI as a cure for past fallacies, with uncompromising assumptions of market-orientation and technology advancements, may stifle the autonomous, diverse and dynamic character of FLI.