In Western Australia soil samples to measure soil‐test phosphorus (P) are collected November to March when soils are usually dry. Most of the soils are hard‐setting when dry and it is difficult to penetrate and collect soil samples to 10 cm using the traditional sampler, which is a 2.5 cm diameter tube that is pushed into the soil by foot. Farmers collect too few soil samples at shallow depths to measure soil‐test P. In Part 1 of this paper, it was shown that soil‐test P can vary markedly for individual soil samples collected from uniform areas in paddocks. Consequently, an adequate number of soil samples needs to be collected and bulked from defined areas to measure soil‐test P. Phopshorus accumulates near the top of the soil of (i) pastures since P fertilisers are applied to the surface of pastures, and (ii) crops sown by minimum (conservation) tillage. Crops are increasingly being sown using conservation tillage methods. Collecting too few soil samples to a shallow depth can result in large errors when estimating the current P status of soils and determining optimum fertiliser application levels. A new rotating blade method of collecting soil samples to the standard 10 cm depth for measuring soil‐test P was compared with the traditional sampler. The new sampler successfully penetrated hard‐setting soils to the standard depth and the soil samples it collected produced similar soil‐test P as the traditional sampler. Consequently, the same soil‐test P calibrations determined using the traditional sampler can also be used for the new sampler. The new sampler is, therefore, recommended for collecting soil samples to measure soil‐test P. It should have application wherever hard‐setting soils pose a problem for soil P testing.