A hybrid deterministic-stochastic algorithm combining the simplex method (SM) and a genetic algorithm (GA) was applied to the problem of extracting the optical and morphological properties of human skin (HSOMPs) from visual reflectance spectroscopy data. The results using the GA-SM hybrid algorithm adopting tournament selection and selecting new sets of HSOMPs were compared with those using other conventional optimization algorithms that have generally been used for the extraction of HSOMPs. Monte-Carlo simulation showed that the suggested GA-SM hybrid algorithm enhanced the stability of the inverse solutions and computational efficiency.