Smallholder land productivity in drylands can be increased by optimizing locally available resources through nutrient enhancement and water conservation. In this study, we investigated the effect of tillage system, organic resource, and chemical nitrogen fertilizer application on maize productivity in a sandy soil in eastern Kenya over four seasons. The objectives were to (1) determine effects of different tillage–organic resource combinations (tillage system and low quantities of organic resources) on soil structure and crop yield, (2) determine optimum organic–inorganic nutrient combinations for arid and semi-arid environments in Kenya, and (3) assess partial nutrient budgets of different soil, water, and nutrient management practices using nutrient inflows and outflows. This experiment, initiated in the short rainy season of 2005, was a split plot design involving tied ridges (under reduced tillage), no-till and conventional tillage, and organic resource (manure and/or crop residue) combinations in the main plots, and chemical nitrogen fertilizer in subplots. Although yield in no-till was lower than conventional and tied ridges during initial two seasons, it achieved higher yields than did the other tillage systems by season 4. Combined application of 1 t ha–1 of crop residue and 1 t ha–1 of manure increased maize yield over sole application of manure at 2 t ha–1 by between 17 and 51% depending on the tillage system for treatments without inorganic N fertilizer. Cumulative nutrients in harvested maize in the four seasons ranged from 57 to 190 kg N ha–1, 11 to 26 kg P ha–1, and 95 to 195 kg K ha–1, representing 23 and 62% of applied N in treatments with and without mineral fertilizer N, respectively, 10% of applied P, and 35% of applied K. There were higher total nutrient NPK uptake treatments under tied ridges than in those under no-till and conventional tillage systems. We conclude that even with modest organic resource application, conservation tillage can be effective in improving crop yield and soil structure and that farmers are better off applying 1 t ha–1 each of crop residue and manure rather than sole manure.