Abstract Methane, as a co-occurring compound, can be found in the majority of coal basins all over the world. The knowledge of methane content is a basic piece of information used in balancing and describing a given coal seam, in conducting safe exploitation – with methane hazards and methane and rock outburst hazards in mind – as well as in less common procedures, such as coal seam fracturing and methane extraction, or storing carbon dioxide in off-balance coal seams. The author of the present paper outlines a thoroughly modified desorbometric method used to analyze the coal-methane system. The research object, in underground conditions, is the full process of releasing methane from coal, which makes it possible to determine gas losses, the desorbable methane content, and the effective diffusion coefficient. This type of research can be conducted in less than 24 hours due to the reduction of the grain fraction of the analyzed bore dust. The thorough research in underground conditions was preceded by a model analysis of the simultaneous impact of the desorbable methane content in coal and the effective diffusion coefficient upon the value of the desorption indensity index. The influence of a grain fraction on the kinetics of methane release from coal was described, and the 0.20–0.25 [mm] grain fraction was selected. The choice of this particular fraction made it possible to register the full process of methane release from coal, in the assumed time. An original methodology of performing underground measurements was presented, together with an instrument used to perform these measurements. The underground research involved measuring the desorbable methane content according to the presented original methodology, which was done simultaneously with an analysis of the coalbed methane content carried out by means of the bore dust method, an analysis of desorption intensity, and an analysis of the technical parameters of coal. The results point to a substantial discrepancy between the desorbable methane content in coal and the coalbed methane content measured by the bore hole method in compliance with the Polish standard. The mean relative percentage error for 17 measurements was 6.5%. The author provided an example concerning using the effective diffusion coefficient in explaining substantial differences in the desorption intensity index for two samples characterized by the similar coalbed methane content.