This study presents an engineering approach for optimizing performance of fluid-film lubricated bearings. Unconstrained nonlinear programming methods, lattice search and simplex method, were used as the optimization schemes to improve the merit of studied bearings with two or more design variables. The analysis of elliptical bearings shows high eccentricity ratio and two large pressure zones for high-speed stability can be obtained by maximizing film pressures in the upper and lower lobes. In this study, lattice method exhibits slightly more efficient search compared with that of simplex method in several two-variable optimum designs. The automatic mesh generation technique used in the pocket-shaped bearing analysis makes the numerical optimization as a flexible design tool. The effect of side flow restrictions on the load-carrying capacity of an optimized pocket-shaped slider bearing is clearly verified. The analysis of the aerostatic bearing explains an example of multi-objective minimization. A similar procedure can be easily adopted to analyze bearings with other profiles, or to maximize user-defined performance using more complicated models. Presented as a Society of Tribologists and Lubrication Engineers Paper at the STLE/ASME Tribology Conference in Orlando, Florida, October 11–13, 1999