We study the well-posedness of a problem for a Petrovskii-parabolic equation with coefficients depending on the space coordinates and with multipoint conditions with respect to the time variable. We establish conditions for the existence and uniqueness of the classical solution of the problem. For the proof of the existence of a solution of the problem, the method of divided differences is used. We prove a metric-type theorem on lower bounds for the small denominators that appear in the construction of the solution.Problems for partial differential equations with multipoint conditions (with respect to both time and space variables) were studied in [1][2][3][4][5][6][7][8][9][10][11][12]. Problems for hyperbolic, parabolic, and typeless equations with multipoint conditions with respect to the selected variable t and certain conditions with respect to the other coordinates (periodicity conditions and Dirichlet-type conditions) were investigated in [1, 6-10]. It was established that these problems are, generally speaking, conditionally well posed, and, in many cases, their solvability is associated with the problem of small denominators.For some classes of higher-order parabolic equations with constant and variable coefficients, local multipoint problems were considered in [9, 10]. The solvability of nonlocal multipoint problems was studied for second-order parabolic equations with variable coefficients in [2,3,5,11] and for systems of parabolic equations in [4,12].In the present paper, which is a continuation of [9], we investigate the correct solvability of a problem for a Petrovskii-parabolic equation with coefficients depending on x = (x 1 , . . . , x p ), multipoint conditions with respect to the time variable, and Dirichlet-type conditions with respect to the space variables. We construct a solution of the problem in the form of a Fourier series in a system of orthogonal functions of x. For the determination of the coefficients u k (t) of the series, we use the fundamental systems of solutions of the corresponding differential equations constructed using divided differences. In contrast to [9], this enables us to avoid small denominators that have the form of the differences of roots of characteristic equations.