A new intelligent orthogonal optimization algorithm for robust design is proposed in order to improve accuracy and efficiency. The next searching direction and searching range of variables are determined by variance ratio after the robust optimization model is firstly calculated by design parameters on orthogonal array. New orthogonal array for further optimization is formed intelligently by analysis of variance ratio. The intelligent orthogonal optimization is performed until error value of each variable is equal to zero or is equivalent, which is the optimal robust solution. Correspondingly, the variable range corresponding to the minimum variance ratio in the orthogonal array in preceding step is the tolerance of the optimal robust solution, which means that there is no need for special tolerance design. This paper takes a cam profile as an example to perform robust design. The simulation results prove that the new intelligent algorithm for robust design has many advantages, such as less calculation time, higher speed, no exiting of prematurity of local circulation and slow convergence of global search.