The article presents the results of a discriminative analysis of the situation of 851 farms that carried out accounting under the supervision of the Institute od Agricultural and Food Economics in 2000 and 2001. The farms were classified on the basis of agricultural income earned in 2001. When the income earned by a farm was lower than the average income (median) the farm in question was included in Class I. When the income of a farm exceeded the average, it was included in Class II. The classes were defined on the basis of a set of financial indicators for 2000 (obtained through a statistical analysis). The group of analysed farms was divided into two parts - a set of farms with typical values of features (fitting into the bracket: average value plus or minus three standard deviations) and a set of the remaining farms whose features had untypical values. For the purpose of discrimination a linear discriminative function was used, which was created on the basis of data coming from 703 farms characterised by the typical value of features, whereas the evaluation was done on the basis of the number of correct classifications for the two sets. The most reliable indicators in forecasting a farm's financial condition in the coming year are: the profitability of sales, the productivity of assets, labour productivity, stock-turn, and the degree of the intensity of organisation of production. All these indicators have a stimulating character, which means that growth in the rates that they represent causes growth in the discriminative value.