A standard recursive identification algorithm for lumped parameter systems is applied in this paper for a class of distributed parameter systems using explicit two time level finite difference models. Each time level has two neighbouring discrete points. The identification is performed in a scalar manner at each time level beginning from the left border of a 2-dimensional spatial-time area. The time level is increased when the computation for the last point at the present level is completed. The nature of the procedure is 1 -dimensional and is recursive for both spatial and time variables. The two neighbouring points at thekth time level influence the values in the discrete model at the (k+ l)th time level. Having obtained the estimates of the parameters at thekth time level we consider in the discrete scalar model the appropriate signals at this level asinputsfor the (k+ l)th level. The proposed procedure of parameter identification is studied via several numerical examples of heat transfer problems with constant parameters.