Abstract. The aim of this study was to classify the cases of calving difficulty using selected data mining methods (Classification and Regression Trees, CART, Chi-square Automatic Interaction Detector, CHAID, and Quick, Unbiased, Efficient, Statistical Trees, QUEST) and generalized linear model (GLZ) and to identify their most important predictors. A total of 1699 records of Polish Holstein-Friesian Black-and-White cows were used. Calving difficulty had three categories (easy, moderate and difficult). Percentages of calvings correctly classified by CART, CHAID, QUEST and GLZ, respectively, were as follows: 60.20, 65.31, 68.88 and 66.33% (easy), 71.36, 69.01, 64.79 and 69.01% (moderate) and 0, 0, 0 and 0% (difficult). The most influential predictors of calving difficulty were the rank of the dam’s sire, calf sex, calving age, previous calving difficulty, lactation number, daily milk yield and average milk yield of the farm. The tree models and GLZ were of moderate quality. None of them could correctly indicate dystocia.