Potassium belongs to the group of alkaline elements, it is the seventh most abundant element in the Earth's crust, and one of the elements that enter into the composition of minerals that form the parent rocks from which the soils are formed. Potassium is found in most soils in relatively large quantities, although its concentrations may vary based on the soil texture. The research is focused on the problem of the status of potassium in different agricultural soils of Eastern Croatian, and the aim of the study was to compare the extraction methods of readily available potassium in the soil with AL-method as the most common method in the Republic of Croatia, to determine the influence of soil type on the suitability of methods for the analysis of easily accessible potassium in the soil, to determine the effects of soil type and physico-chemical properties on the fixation of potassium, and to create a computer model that estimates fixation of potassium in different soils. Exchangeable acidity (pH KCl) of investigated soils ranged from 3.27 to 7.91, organic matter content ranged from 0.83 to 11.80%, AL-P ranged from 0.58 to 171.88mg/100g , AL-K from 6.88 to 63,35 mg 100g, Hy 0.22 to 8.40, and carbonate content from 0.84 to 16.13%, which means that the samples were very heterogeneous. All soil types have statistically significant difference in the dynamics of potassium, where substantially highest fraction at all soil types was total potassium fraction, followed by potassium fraction that is incorporated in the crystal structure, fraction of low available potassium, and finally fraction of the readily available potassium. Comparison of analyzed extraction methods of potassium from the soil is of high precision, and most reliable comparisons are of AL methods with AAEDTA, followed by a: AA, BaCl2, ECF, Bray, EUF1 and ECF 2 method. Regression models are very simple and applicable since they can provide new information based on the concentration of potassium extracted AL method. The accuracy of the model significantly increases by using additional soil properties data, mainly the pH reaction and organic matter content. Additional precision of the model can increase by adding some additional soil properties data such as CEC and proportion of potassium in CEC. The developed model is necessary to be validated on the data sets of known analytically determined values from field experiments.