An analytical numerical model to optimize the shape of concave surface texture for the achievement of low friction in reciprocating sliding motion has been developed. The model uses: (i) average Reynolds equation to evaluate friction coefficient and (ii) genetic algorithm (GA) to optimize and obtain several preferable texture shapes. Analysis of distribution contour maps of hydrodynamic pressure gives the possible mechanisms involved. Moreover, experimental comparisons of tribological performances between the optimized and the circular textures were conducted to verify the simulation results. It is shown that surface textures of the elliptical and fusiform shapes can effectively enhance the load bearing capacity and reduce the friction coefficient compared with circular textures. The increase in hydrodynamic pressure for these optimized texture shapes is considered to be the major mechanism responsible for improving their tribological performance. Experimental results confirm that the elliptical-shaped textures have preferable tribological behaviors of low friction coefficient under the operating condition of light load.