Although the application of small quantities of nitrogen (N) fertiliser has improved cereal yields on low-input farms in semi-arid Zimbabwe, the practice is reported to be laborious and time-consuming by farmers. In an effort to make micro-dosing less labour-intensive and more precise, an ammonium nitrate (AN) tablet, the equivalent of a micro-dose of prill AN (28 kg N ha−1) applied per maize plant, was developed by International Crops Research Institute for the Semi-Arid Tropics in collaboration with Agri-Seeds, Zimbabwe. This study characterized the physical stability, chemical (N% and solubility) and agronomic performance of AN tablets compared with prill AN. Only 10% of tablets broke when dropped from 2 m, showing that they are physically stable and can handle rough treatment. The N content in the tablets (33.3%) was comparable to that in prill AN (34.6%). However, the tablet formulation took twice as long to dissolve than prill AN when placed on a wet soil. Despite this difference in solubility, simple leaching column experiments suggest that less than 2% of the total AN applied was lost due to leaching. Agronomic trials were superimposed on the paired-plot demonstrations used to promote micro-dosing and the conservation agriculture tillage technique of planting basins from 2005 to 2008. Each tillage (plough and basins) plot was subdivided into three sub-plots on which no AN, prill AN and tableted AN treatments were superimposed. Maize was planted and management of plots was left to farmers. Micro-dosing with either prill or tableted AN significantly (p < 0.001) increased maize grain yield by over 40% in all seasons for planting basins. However, on the ploughed plot there was no yield benefit to using either AN formulation in the season with the lowest rainfall (2006–2007). There was no significant difference in grain yield and agronomic N use efficiency between prill and tableted AN formulations except for the 2005–2006 season in planting basins. During this season, in planting basins, tableted AN had significantly (p < 0.001) higher rainwater productivity than prill AN, which translated into greater grain yield. In addition, the maximum benefit to micro-dosing was observed to accrue when combined with water harvesting techniques such as planting basins. An observation supported by the host farmers, who in the second and third seasons chose to apply available basal soil fertility amendments to the basin plots over the flat plots. Thus, AN tablets if available at an affordable price can be used by smallholder farmers to more precisely apply N fertiliser. Future work should focus on the labour issues of micro-dosing, and making cost-effective tablets available to resource-poor farmers and also addressing other limiting soil nutrients.