Fertilizer placement may cause non‐uniform nutrient distribution in the soil, making it difficult to determine whole‐field fertility by traditional sampling strategies. Our objectives were to determine P and K distribution after repeated applications in no‐till and strip‐till soils and to develop improved sampling procedures to estimate soil P and K levels on a corn (Zea mays L.) and soybean [Glycine max (L.) Merr.] rotation with crops planted at 76‐cm row spacing. Three trials near Pesotum, IL, received blends of 0–0, 22–42, 33–62, 44–83, 55–104, 66–125, and 77–145 kg P–K ha−1 in fall 2007 and 2009 before corn planting. Applications were broadcast‐applied in no‐till (NTBC) and strip‐till (STBC) and deep‐banded in strip‐till (STDB) 15 cm below the surface in the crop row (IR) using real‐time kinematic (RTK) satellite navigation. Every year soil P and K was measured at 10‐cm increments to a 30‐cm depth at 0, 19, 38, and 57 cm from the IR. Subsurface banding reduced P and K levels in the surface and increased them at the point of application, or deeper with the highest rate, while broadcast applications increased surface levels. Soil‐surface K levels were greater at IR likely because of K leaching from senescing standing crops. Soil‐test results indicated no need to adjust fertilizer rate based on tillage or fertilizer placement. A sampling ratio of 1:3 IR to between the crop rows (BR) seemed adequate to estimate soil fertility across a wide range of P‐ and K‐fertilizer rates and soil test levels where the location of the fertilizer band or planting row is maintained constant.