The basis of modelling real estate unit values is information on prices and qualities of real estates as subject of market turnover or on market values of representative real estates and their attributes. Such a data set constitutes a representative basis of real estates for analysing the market, meeting all criteria of multidimensional random variable. Among different non-linear functions of many variables, in modelling of a real estate unit price or value, multiplicative exponential function is chosen to be examined in relation to the particular attributes. Exponential form of the function assures positive values of real estates and it permits to describe their variability as a monotone function. The model of real estates unit values in form of multiplicative exponential function was analysed as follow $c=B_{0} \cdot B_{1}^{x_1} \cdot B_{2}^{x_2}...B_{m}^{x_m}$, where: $c$ - real estate unit price or unit value, $x_1, x_2,..., x_m$ - real estate attributes, including the attribute of »transaction time«, $B_j$ - estimated model factors, $B_0$ - real estate unit value, for zero of all attributes. Estimation of model parameters may be done in many ways. In the present paper, the run system for Markow method (in form of least squares weighing method) will be submitted. Verification of estimated model proceeds as follows: 1. investigation of model acceptability regarding the values of factors variability as well as the convergence, 2. analysis of model factors significance, 3. analysis of random components symmetry.
