The most important step in packaging printing is determining the spot color blending rate. Currently, in most processes, spot colors are produced based on the operator's experience. Accordingly, the process is managed inefficiently and unnecessary resources are consumed. To solve this problem, we intend to establish an artificial intelligence model that can predict the color coordinates of spot colors. An optimal algorithm is constructed to predict the color coordinates of the spot color. Based on the stacking ensemble, four models are firstly generated, and secondly a single model that receives outputs of the generated four models is generated to develop an algorithm. As a result of learning the developed model by gradually increasing the number of data to check the performance of the model according to the number of data, it can be seen that the performance of the model increases as the number of data increases. In addition, as a result of training the model using all the data, it can be seen that it predicts the CIE L*, a*, and b* of the output well. In conclusion, beyond the mathematical model that can predict color coordinates well only in a specific structure, the artificial intelligence model implemented through data learning predicts the color coordinates of spot colors considerably. This suggests that artificial intelligence can also be used in color-related fields. In this study, effective process management can be carried out by reducing time and minimizing resources by applying the learned artificial intelligence model to the process. In addition, it is expected that artificial intelligence that can accurately predict colors by acquiring a lot of data in the future will be developed.