Grinding induces strong residual stress gradients in the workpiece, espescially in brittle materials. Stress gradients are of interest to optimize production processes. The algorithm discussed in this paper chooses spline functions as model functions to recover the stress profile from X-ray diffraction data. Simulated data are used to determine a good choice for the degree of the polynomials, the number of sections, and the depth in which the stress profile is regarded. The algorithm shows a good reevaluation of gradients near the surface even for noisy data. The difference between the original function and the reevaluated function increases with increasing depth.