Abstract: To achieve an optimal balance between the reliability performance and R&D cost control of CNC machine tools, a reliability allocation method integrating the normal cloud model, the criteria importance through intercriteria correlation (CRITIC) algorithm, and the visekriterijumska optimizacija I kompromisno resenje (VIKOR) algorithm is proposed. The normal cloud model is employed to process expert group decisions, retaining the fuzziness and randomness of the evaluations. Based on the volatility and conflict of cloud feature data, the CRITIC algorithm is improved to achieve the objective weighting of risk factors. Furthermore, the VIKOR compromise ranking mechanism is introduced to capture the shortboard effect of the machine tool system. Taking a dual-spindle machining center as the research object for calculation and analysis, the results indicate that compared with the traditional allocation method based on intuitionistic trapezoidal fuzzy numbers and analytic hierarchy process (AHP), the mean time between failures (MTBF) allocation target for the spindle system with harsh working conditions is further increased from 14 543 h to 23 976 h, while the MTBF target for the technically mature protection system is reasonably reduced from 30 303 h to 21 339 h. This allocation strategy overcomes the drawbacks of averaging allocation, scientifically realizes the strict target elevation for core shortboard links and moderate target reduction and burden relief for mature auxiliary links, and provides a quantitative decision-making basis for the precise decomposition of reliability indicators in high-end machine tools.