A method for the generation of bicubic spline functions is presented in this paper. From this method it becomes apparent that these functions derive their potential strength in accurate and reliable representation of two‐dimensional data by maintaining continuity of the variable and its slope and curvature throughout the area of observation. The results obtained by computing horizontal and vertical derivatives with model and field data illustrate the exceptional accuracy achieved with spline functions. The piecewise cubic polynomial functions expressing observed data analytically in space are used to estimate amplitude and phase spectra of magnetic anomalies. At relatively long wavelengths the amplitude spectrum thus calculated displays remarkable similarity with the true spectrum and is found to be superior to that obtained with two‐dimensional Fourier series expansion. A cubic spline method is also presented for computing values of an observed variable at equispaced points along two orthogonal directions with the help of irregularly distributed data. The interpolation technique applied to field data shows high resolution by maintaining the separation of neighboring anomalies and the small‐scale features. The shapes, peaks, and troughs of both large and small amplitude anomalies are faithfully reproduced. The gradients of the magnetic field do not undergo any appreciable distortion. It can thus be concluded that cubic splines are a reliable and accurate method of interpolation.