Optimized surface mine layouts are used to extract mineable reserves with minimum waste under economic geological, geotechnical, and property boundary constraints. Surface mine design and optimization algorithms are limited in dealing with the random field properties of these layouts, resulting in suboptimal results. Database changes also require complete rerun of these algorithms, resulting in long CPU times with no allowance for incorporating operating strategies. In this study, the authors develop a computational intelligent (Cl) algorithm to solve these problems. The Cl algorithm combines the stochastic models of ore reserves and commodity prices to generate economic block and target values. The error back-propagation algorithm is used to train feed-forward neural networks for block pattern recognition and partitioning based on the target values. The Cl algorithm is used to optimize Section SBHP 860001 of a surface mine layout, and the results are compared with that from the 2-D Lerchs-Grossmann's algorithm.