The main objective of the paper was to propose and evaluate the performance of a regional approach to estimate CN values and to test the impact of different initial abstraction ratios. The curve number (CN) was analyzed for five Slovak and five Polish catchments situated in the Carpathian Mountains. The L-moment based method of Hosking and Wallis and the ANOVA test were combined to delineate the area in two homogenous regions of catchments with similar CN values. The optimization condition enabled the choice of the initial abstraction ratio, which provided the smallest discrepancy between the tabulated and estimated CNs and the antecedent runoff conditions. The homogeneity in the CN within the regions of four Slovak and four Polish catchments was revealed. Finally, the regional CN was proposed to be at the 50% quantile of the regional theoretical distribution function estimated from all the CNs in the region. The approach is applied in a group of Slovak and Polish catchments with physiographic conditions representative for the Carpathian region. The main benefit of introducing a common regional CN is the opportunity to apply this procedure in catchments of similar soil-physiographic characteristics and to verify the existing tabulated CN. The paper could give rise to an alternative way of estimating the CN values in forested catchments and catchments with a lack of data or without observations.