It was assumed that the production of alfalfa in soils naturally poor in available nutrients, such as potassium (K) and calcium (Ca), depends on the use of fertilizers. This hypothesis was validated in an experiment with an alfalfa–grass mixture carried out in 2012, 2013 and 2014 on soil formed from loamy sand that had a low content of available Ca and K. The two-factor experiment consisted of two levels of applied gypsum as a source of Ca (0, 500 kg ha−1) and five levels of PK fertilizers (absolute control, P60K0, P60K30, P60K60 and P60K120). The total yield of the sward was determined by the main seasons of alfalfa–grass sward use. Gypsum application increased the yield by 1.0 t ha−1. The highest yield of 14.9 t ha−1 was obtained on the plot fertilized with P60K120. Based on the nutrient content in the sward, it was shown that the main yield predictor was the content of K in the first cut of sward use. The reliable yield predictors, based on the total accumulation of nutrients in the sward, turned out to be K, Mg and Fe. The nutritional quality of the alfalfa–grass fodder, based on the K/Ca + Mg ratio, depended mainly on the season of the sward use, which was substantially deteriorated by the K fertilizer. Gypsum did not control this process. The productivity of the nutrients taken up by the sward depended on the accumulated K. Its yield-forming effect was significantly limited by manganese deficiency. The use of gypsum positively affected the uptake of micronutrients, consequently increasing their unit productivity, especially of manganese. Optimization of the production of alfalfa–grass mixtures in soils poor in basic nutrients requires micronutrients to be taken into account. Their uptake by plants can be limited by high doses of basic fertilizers.