The direct application of rock phosphate (RP) is known to be a good alternative to water-soluble P fertilizers. However, there are no models to predict amounts of RP required. An incubation study was conducted to develop an algorithm for estimating RP requirements in West African soils. The dissolved RP was measured by the changes in 0.5 M NaOH-P over time. The parameters of the equation fitted to the dissolution curves were regressed against soil properties to identify and quantify the soil properties that best predicted RP dissolution. The amount of RP dissolved could be described by an equation of the form: Y=A-(B*e^-c*time), where Y=amount of RP dissolved; A=78.1627+(30.62149*KCl-extractable acidity)–(0.63335*CaSat) [residual mean square, RMS=0.16132]; B=-16.58786–(21.74709*KCl-extractable acidity)+(4.07675*ECEC) [RMS=0.01463] and c=0.142. The changes in Bray1- and Olsen-P during the incubation period, the initial soil and plant critical P levels, the clay content and the dissolution equation were used to predict the amounts of RP needed. A study was conducted in Bambey, Senegal to field test the predictions made by the algorithm. Treble super phosphate (TSP), Gafsa (GRP), Tilemsi (TRP) and Taiba (TaRP) RP were applied at 0, ¼X, ½X, 1X and 2X, where 1X was equivalent to 113, 2734, 5361.33 and 1301.33 kg ha^-1 of TSP, TRP, TaRP and GRP, reflecting the different RP qualities. The maximum millet dry biomass yields were the same for TSP, TRP and TaRP but less for GRP. The maximum biomass yields were achieved with the application of 86.55 kg TSP ha^-1, 673.5459 kg TRP ha^-1, 3415.8205 kg TaRP ha^-1 and 652.08 kg GRP ha^-1. The proposed algorithm overestimated the amounts of RP needed to supply crop P requirements. Would the removal of the calcium dissolved from the RP in field conditions enable more dissolution to occur and explain the overestimation of the algorithm predictions?