This paper researches the real valued random fields, that are homogeneous with respect to time and homogeneous isotropic with respect to spatial variables. An analogue of the Kotelnikov-Shannon theorem for random fields with a bounded spectrum is presented. Models for such random fields by partial sums of series are constructed. Some estimates for the mean square approximation of a random field by its models are obtained. Statistical simulation procedures of realizations of a random field with Gaussian distribution are constructed. The using of these theorems, models and procedures are demonstrated through applications to generate by means of computer adequate realizations of Gaussian random field with some wide-known examples of covariance functions. Spectral analysis of generated noise is considered.