Water-extractable phosphorus in soil has been interpreted as a rough index of soil phosphorus fertility, but it might be inadequate as an index to predict phosphorus uptake by crops, propose detailed recommendations for fertilizer application rates, or both. However, a relationship has been reported between the amount of water-extracted phosphorus at equilibrium (Q) and water/soil ratio (w/s ratio, W): 1/Q = b/W + a, where a and b are constants. Using this equation, one could predict the maximum extractable phosphorus as 1/a, corresponding to theoretical extraction with an infinite volume of water, and this value could serve as a capacity factor. We are developing a new non-shaking water extraction (NSWE) method suitable for on-farm applications, and we are interested in whether the 1/Q versus 1/W relationship also holds true for this method. Because NSWE is a closed system, unlike infinite sink type extraction methods such as the ion exchange resin method, phosphorus extraction was anticipated to reach the maximum in a relatively short period, and several extraction periods (up to 48 h) were tested. However, the time required to reach equilibration differed among various w/s ratios and the apparent equilibrium was not experimentally observed in some cases. Therefore, we tested first- and second-order kinetic models, in which Q values can be identified, in order to predict Q for NSWE. Although the first-order model needed slight modification, namely assuming the existence of a very quickly extractable component, the model predicted the temporal change fairly well, and the predicted Q values acceptably fit the 1/Q versus 1/W relationship. The second-order model was slightly less precise in predicting the temporal change. This finding suggests that the maximum extractable phosphorus from NSWE also can be predicted by the 1/Q versus 1/W relationship.