Carbon dioxide (CO₂) flooding is an effective method of enhanced oil recovery (EOR) that has become one of the most important EOR processes. One of the key factors in the design of a CO₂ injection project is the minimum miscibility pressure (MMP), whereas local sweeping efficiency during gas injection is dependent on the MMP. There are many empirical correlation analyses for the MMP calculation. However, these analyses focus on the molecular weight of the C₅₊ or C₇₊ fraction, and do not emphasize the effects of various components on MMP. Our study aims to develop an improved CO₂–oil MMP correlation analysis that includes parameters such as reservoir temperature and various oil mole fractions. Here, correlation analysis was performed to define the influence of various components on the MMP using various data from 45 oilfields which have experimental CO₂–oil MMP and oil compositions readily available. Thirty of the data sets were used to develop an improved correlation, and the other 15 data sets were used to verify the correlation. It was found that the mole fraction of C₃ and C₆ were the main factors that affected MMP. There was a good quadratic polynomial relationship between the mole fraction of C₃ and MMP, and the relationship also existed between the mole fraction of C₆ and MMP. The results do not include the molecular weight of the C₅₊ or C₇₊ fraction like other common correlations. Nine popular correlations were then used to also predict the MMP, and the comparison showed that the improved CO₂–oil MMP correlation defined here was a better estimate. The correlation was then used in Dongshisi and Fuyu oilfields to assess EOR potential, the results also indicated that MMP increased over the course of the CO₂ flooding process. This increase shows that it would be more difficult to achieve a mixed phase between crude oil and CO₂, therefore the oil recovery would be difficult to further improve towards the end of injection.