【Objective】This study was to explore effect of long-term potassium application on soil potassium supply capacity, so as to provide certain scientific basis for rationally potassium application and continuing research on development of agricultural soil.【Method】From a long term potassium fertilization experiment on a tobacco-rape rotation system, which was designed to have 0,187.5 or 375 kg·hm^-2 K₂O applied to flue-cured tobacco and 36 kg·hm^-2 to rape, soil samples were collected for analysis of apparent balance and Quantity-Intensity (Q/I) of soil potassium, and relevant parameters, i.e. K equilibrium activity ratio (AR_e), soil labile K (K_L), non-specific labile K (-ΔK₀) and specific labile K (K_X), potential buffering capacity (PBC), K and Ca+Mg exchange free energy (-ΔG), and soil properties, like readily available K (K_av), slowly available K (K_nex), K saturation degree (K_S), clay minerals components and so on, and moreover, relationships between Q/I parameters and soil properties in the flue-cured tobacco-rape cropping system.【Result】 Results show that after 20 years of the stationary experiment, ARe varied in the range of 0.38×10^-3~8.91×10^-3（mol·L^-1）^0.5, -ΔK0 of 0.02~0.55 cmol·kg^-1, KL of 1.61~2.16 cmol·kg^-1, K_X of 1.06-2.14 cmol·kg^-1, PBC of 6.20~7.76 cmol·kg^-1(mol·L^-1)^-0.5, and -ΔG of 11.70~19.72 kJ·mol^-1. Obviously K application enhanced both AR_E and -ΔK₀, but reduced K_L, K_X, PBC, and -ΔG. K_av, K_nex, K_S and K_ab were extre-significantly and positively correlated with AR_e and -ΔK₀, but negatively with K_L, K_X, PBC and -ΔG. Significant relationships were observed between the parameters of the Q/I curve of soil K, and these parameters could be integrated into a principal component index, which was significantly lower in K application treatments than in no K application treatments. Soil potassium was retained mainly interlaminarly in clay mineral crystals and lower in bio-availability in the soils with no K applied, but on joint position interface of the clay mineral crystals and higher in bio-availability in the soils applied with K. Hence, the soils with no K applied were higher in vermiculite content, but lower in illite content than the soils with K applied.【Conclusion】All the Q/I curve parameters, such as AR_e, -ΔK₀, K_L, K_X, PBC,-ΔG, can be used as indicators for evaluation of soil potassium supply capacity. Apparent balance of soil potassium is a major factor driving changes in K Q/I curve shape and differentiation of its parameters, variation of potassium supply capacity and evolution of soil minerals. The higher the K application rate, the higher the soil potassium surplus and the higher the soil supply capacity. In the soils under long term potassium deficiency, mineral potassium transforms into specific labile K at a higher rate than specific labile K does into non-specific labile K. Long-term potassium deficiency causes transformation of illite into vermiculite.