In this work, an iterative scheme using cubic splines with defect two is considered for a boundary value problem for neutral delay linear differential-difference equations. The conditions for the boundary value problem solution existence for various classes of differentialdifference equations were considered in [1-3]. Let us consider the following boundary value problem y′′ (x) = n i=0  ai (x) y (x − τi (x)) + bi (x) y′ (x − τi (x))+ (1) +ci (x) y′′ (x − τi (x))  + f ((x) , y(p) (x) = φ(p) (x), p = 0, 1, 2, x ∈ [a∗; a], y(b) = γ, (2) where τ0 (x) = 0 and τi (x), i = 1,n are continuous nonnegative functions defined on [a, b], φ (x) is a continuously differentiable function given on [a∗; a], γ ∈ R, a∗ = min 0≤i<n  inf x∈[a;b] (x − τi (x))  . We introduce the sets of points determined by the delays τ1 (x) , . . . , τn (x): Ei1 =  xj ∈ [a, b] : xj − τi (xj) = a, j = 1, 2, . . .  , Ei2 =  xj ∈ [a, b] : x0 = a, xj+1 − τi (xj+1) = xj, j = 0, 1, 2, . . .  , E2 = n i=1 (Ei1 ∪ Ei2) . A function y = y (x) is called a solution of the problem (1)-(2) if it satisfies the equation (1) on [a; b] (with the possible exception of the set E2) and boundary conditions (2). We will set an irregular grid Δ = {a = x0 < x1 < .. . < xm = b} on [a; b] such that E2 ⊂ Δ.For finding an approximate solution of the boundary value problem (1)-(2) a computational scheme in the form of a sequence of cubic splines with defect 2 on the grid Δ is proposed and substantiated [4].