The three-year field experiment (2015–2017) with graded doses of magnesium (Mg) and sulphur (S) was carried out at the Humpolec experimental station (49.5546239N, 15.3485489E; Czech Republic). The interactions between boron (B), Mg and S in the soil were studied. No boron was applied into soils. Contents of B, S and Mg in the soil were determined by the Mehlich 3 and NH<sub>4</sub> acetate methods. The crop rotation was: spring barley-oilseed rape-winter wheat. Three Kieserite doses (S and Mg fertiliser) were applied. Sulphur treatments were 10-20-40 kg S/ha to cereals and 20-40-80 kg S/ha to oilseed rape. The doses of Mg were: 13-26-52 kg Mg/ha to cereals and 26-52-104 kg Mg/ha to oilseed rape. A significant gradual decrease of B-Mehlich 3 was observed under Kieserite treatments during the experiment (from 1.24 mg B/kg in control in the 1<sup>st</sup> year to 0.92 mg B/kg in the 3<sup>rd</sup> year). On the contrary, B-NH<sub>4</sub> acetate contents in soils remained similar during 2015–2017 in control soils (0.33–0.39 mg B/kg) and significantly decreased under Kieserite treatments, namely by 55–57% in 2016 and by 43–48% in 2017. A significant decrease of B content in soils was noted since the second year of experiment after oilseed rape. The boron contents in soils were affected in several ways – by adsorption of B on magnesium oxides and other substances, exchange with SO<sub>4</sub><sup>2–</sup> anions and possible leaching, and also by the uptake by grown crops, mainly oilseed rape.