These residues were defined after creating the geostatistical model for the group of chosen real estates in the city of Olsztyn ( northern-eastern Poland). The following procedure was used: ï Initial verification of the attributes and the selection of factors, ï Defining the structure of the mathematical model, ï Defining the interaction of parts of the model, ï Elimination of inadequate incidents, ï The final form of statistical model. // The objects of the analysis were the prices of land real estates (not developed) in 1997-2003. The amount of 821 land real estate properties were studied and 25 attributes were created. Good measures of model adaptation were achieved (R2 > 0,60). In the result of detailed proceeding, the individual features of the analysed model were defined and the map of residues from the regression for Olsztyn was created. The models built on the residues can be used in the mass real estate assessment, in urban real estate management and in the process of developing areas between the urban and rural areas (fringe zones). The algorithms proposed in the paper have universal character and can be successfully used in large and middle-size cities.