The reaction of two genotypes to the same nitrogen–phosphorus–potassium (NPK) fertilisation policy with different magnesium (Mg) fertilisation regimes was studied in two adjacent factorial-design experiments on a mineral soil in Indonesia. The different Mg-fertilisation regimes were chosen after observing contrasted foliar Mg contents in the two different trials. The first experiment (ALCP61—planted with (DA5D×DA3D) × LM2T material with a low foliar Mg signature) received dolomite at a higher rate than in the second experiment (ALCP62—planted with a high foliar Mg signature with (DA5DxDA3D) × LM311P material). After 10 years of this different fertilisation regime, yield was evaluated as well as nutrition of oil palm through foliar and rachis analysis, complemented with soil analysis. The high Mg foliar content in the ALCP62 trial misled us to thinking that Mg fertilisation was not needed. Indeed, the limited supply of dolomite in ALCP62 induced changes in the cation equilibrium which was prejudicial to the nutrient balance in the leaf, as well as in the soil, especially when large amounts of potassium chloride were applied. It acidified the soil, decreased the exchangeable calcium and consequently the available P and total P, leading to an imbalanced palm nutrition which affected production. It allows us to conclude that the differences in foliar contents (Mg in this study) are specific to certain genotypes and need to be considered as such. Therefore, it is important to improve our knowledge in the nutrition × genotype interactions to obtain reliable indicators for responsible fertiliser management.