To solve the complex control problems of multiple key quality characteristics of NC machine tools, based on meta-action theory and two-dimensional House of Quality(HOQ), a three-dimensional House of Quality (THOQ) model with the influencing factors dimension, the meta-action units dimension and the key quality characteristics dimension was established. The action layer factors affecting the key quality characteristics were screened by Ant Colony Clustering Algorithm, and combined with Evidence Reasoning Recursive Theory(ERRT), the autocorrelation matrix of the factors, the correlation matrix between the key quality characteristics and the influencing factors of each meta action unit and performance coupling matrix of meta-action units were established. Based on quality control cost constraint and development time constraint, an optimization model with key quality characteristics fluctuation degree as objective function was established, and the optimal value of the influencing factors were obtained, so that the fluctuation value of the key quality characteristics was minimized. Finally, an example was given to optimize the indexing table motion of a NC machine tool, and the rationality and validity of the model were verified by experiments.