In this article the integral representation for bounded even positive functions $k(x)$\linebreak $\left(x\in \mathbb{R}^\infty=\mathbb{R}\times\mathbb{R}\times\dots \right)$ is proved. We understand the positive the positive definite in the integral sense with integration respects to measure $d\theta(x)= p(x_1)dx_1\otimes p(x_2)dx_2\otimes \dots$ $\left(p(x)=\sqrt{\frac{1}{\pi}}e^{-x^2} \right)$. This integral representation has the form
\begin{equation}\label{ovl1.0}
k(x)=\int\limits_{l_2^+} {\rm Cos}\,\lambda_ix_id\rho(\lambda)
\end{equation}
Equality stands to reason for almost all $x\in \mathbb{R}^\infty$. $l_2^+$ space consists of those vectors $\lambda\in\mathbb{R}^\infty_+=\mathbb{R}^1_+\times \mathbb{R}^1_+\times\dots\left| \sum\limits_{i=1}^\infty \lambda_i^2 <\infty\right.$. Conversely, every integral of form [1] is bounded by even positively definite function $k(x)$ $x\in\mathbb{R}^\infty$.

As a result, from this theorem we shall get generalization of theorem of R.~A.~Minlos--V.~V.~Sazonov \cite{lov2,lov3} in case of bounded even positively definite functions $k(x)$ $(x\in H)$, which are continuous in $O$ in $j$"=topology.

[1] Berezansky Yu. M., Gali I. M. Positive definite functions infinitely many variables in the layer. Ukr. Math. Jour. 1972, 24 (4), 435–463 (in Russian).

[2] Minlos R. A. Generalized random processes and their continuation to the measure. Tr. Mosk. Mat. O-va 1959, 8 (in Russian).

[3] Sazonov V. V. Remark on characteristic functionals. Probability theory and its application 1958, 3 (2) (in Russian).

[4] Shilow G. E., Fan Dyk Tin. Integral, measure and derivative on linear spaces. M, «Science», 1967 (in Russian).

[5] Gichman I. I., Skorokhod A. V. The theory of random processes, M, «Science», 1971 (in Russian).

[6] Berezansky Yu. M., Kondratiev Yu. G. Spectral methods in infinite-dimensional analysis. Kiev, «Naukova dumka», 1988 (in Russian).