Abstract: An optimization method based on blind number was presented with blind number expressing the uncertain variables in optimization and in combination with the traditional optimization algorithm which based on the real variables. In the microscopic point of view, the relation between design variables and optimization parameters was analyzed and described, and the blind number solution was provided. The blind number solution not only provided the values for design variables, but also provided the evaluation of reliability of the optimization object in optimal condition when the design variables had different values.