The precise knowledge of soil nitrogen (N) and phosphorus (P) availability to crops remains an essential factor in sustaining production systems. Anionic exchange membranes (AEMs) have been successfully used to estimate the availability of soil N and P in conventional farming. This technique is not common in precision farming. The objectives of this study were: i) to assess the effects of soil texture, sampling time, and distance (distance from the corn row position) on soil N and P desorbed from AEMs (AEM-N and AEM-P, respectively) and; ii) to characterize the magnitude (coefficient of variation: CV) and structure (semivariogram parameters) of short range spatial and temporal variation of soil AEM-N and AEM-P. The study was conducted on a commercial farm located in the St-Lawrence Lowlands (Quebec, Canada) in 2003. The AEMs were installed at six sampling points covering three soil types: sandy loam, clay loam, and clay. For each point, 64 AEMs (4 × 16 inter-row locations) were positioned on a 3.0 m × 2.25 m grid at three sampling dates: before seeding, at the 6-leaf stage, and at the silk stage. A significant interaction between soil texture, time, and distance was only observed on AEM-N. Coefficient of variation varied from 50 to 138% for AEM-N, and from 52 to 263% for AEM-P. For both elements, a higher CV was observed when AEMs were placed on the row than at 15 and 30 cm. Weak spatial structures were also evident from semi-variogram parameters, indicating a predominance of random factors on measured AEM-N and AEM-P. The short-range spatial variation measuring NO₃ and PO₄ using AEMs was generally too great for reliable sampling strategy design.