The goal of this study was to investigate the effect of missing data in the covariate variable on the amount of lost information in the cure mixture survival analysis model. For this purpose, data from patients of a cardiovascular cohort in Sulaimani Hospital were used. In the case of missing completely at random, in six situations, include full data and 10% to 50% missing data, AIC and BIC values of the models were compared. The results showed that the behavior of these two indicators was the same in comparing the models. Also, the choice of the appropriate imputation method for cured mixture models depends on the rate of missing data. In 10% of missing, due to the low number of missing individuals, the simple method such as mean was the best method. But for higher percentages, hot deck and regression methods performed better. Also, due to the skewed distribution of patients' age data, in cases with a missing of more than 40%, the median has shown better performance.