Appropriate targeting of soil fertility technologies across temporal and spatial scales has remained a major challenge for research and development agencies which primarily aim to enhance food security and increase farm incomes in smallholder farming systems of Southern Africa. Despite notable advances in soil process research, understanding of the factors driving spatial and temporal variability of soil fertility at field and farm scales has largely been elusive. Empirical information upon which appropriate management decision tools can be drawn is generally lacking. This study was aimed at developing a understanding of the major causes of within field/farm soil fertility gradients and quantifying their potential influence on mineral fertilizer use efficiency under the predominantly maize-based systems of Zimbabwe, using farmer participatory research approaches. The study was conducted over three seasons (2002-05) on 120 field sites under high, medium and low rainfall regions, all of which are dominated by granitic sandy soils (Arenosols/Lixisols) commonly known for their low agricultural potential (< 100 g clay kg^-1; < 10 g C kg^-1). The soils typify most of those in the smallholder farming systems of Zimbabwe. Three distinct forms of fertility gradients, namely transitory, temporary and permanent, were defined, with the former two clearly driven by organic matter management practices and differentiated by farmer resource endowments. The temporary forms of fertility gradients were, to a greater extent, reflected in levels of soil organic C (SOC), and apparently provide opportunities for innovative manipulation by farmers to increase overall farm productivity. The soils evidently had limited capacity to stabilize C, with SOC values ranging from 3 to 8.3 g C kg^-1 but the small differences among field or field sections had a significant influence on mineral fertilizer response by maize. There were three distinct SOC categories across the field sites, over which crop response to mineral fertilizer application was significantly different. Little or no crop response was recorded in fields with <4.6 g C kg^-1 SOC regardless of management practice, and maize yields were invariably <0.9 Mg ha^-1. The highest variability in fertilizer responses was in fields with SOC levels between 4.6 and 6.5 g C kg^-1. Further yield increases were achieved at SOC > 6.5, but regular application of organic inputs at rates exceeding 20 Mg ha^-1 did not increase SOC beyond 8.5 g C kg^-1, suggesting that the goal for organic matter management in these systems should not be that of building high SOC stocks but to maintain the critical levels necessary for increased nutrient use efficiency. These preliminary findings provide an important step in advising national extension agencies and non-governmental organizations on appropriate domains for targeting technologies in order to enhance uptake by smallholder farmers.