The results of a joint project of institutes in the Czech Republic, Hungary, Poland, and the Netherlands on calibration of the universal 0.01M calcium chloride (CaCl2) soil testing procedure for potassium (K) are presented. In soil samples collected in the participating countries, the conventional extraction procedures were compared with the 0.01M CaCl2procedure. Linear and multifactorial regression analysis were applied to quantify the relations of K in the 0.01M CaCl2extract and in the different conventional extractants. With the relatively weak unbuffered extractant 0.01M CaCl2less K is extracted from the soils compared to the relatively strong conventional extraction solutions [Mehlich‐2, Egner‐Riehm, ammonium lactate, and hydrochloric acid (HCl)]. In general the relative amount of K extracted from a soil sample using 0.01M CaCl2is lower the higher the clay and organic matter content of the soil is. The explained variance (r2) between the 0.01M CaCl2procedure and conventional methods varied from 0.53 to 0.96. Relationships improved when clay content or organic matter content were taken into account.