Abstarct: In the underground stopes, the contamination of the ore with wastes or the materials lower than cut-off grade is referred to as dilution. The dilution is important especially in cut and fill stoping method due to weak hanging-wall and footwall. Reduction of dilution is the ways to enhance the quality of products, reduce the associated costs and hence to increase the profit. The aim of this research is to provide a model for assessing and predicting the dilution in cut and fill stoping method. For this purpose, at first all the parameters contributing to unplanned dilution in underground stopes and specifically in the cut-and-fill stoping method were identified. Then, all the parameters were weighed with the use of the Fuzzy-Delphi Analytical Hierarchy Process (FDAHP). After that, the most effective parameters were selected. Then, in order to present a novel classification system for an unplanned dilution assessment, a new index called stope unplanned dilution index (SUDI) was introduced. This index which varies between 10 to 100 represents the extent to which a cut-and-fill stope is susceptible to unplanned dilution. Next, the SUDI index was used for 10 stopes in different parts of Venarch Manganese Mines in Iran. As follows, the cavity monitoring system was implemented in these stopes in order to determine actual dilution. Then, an equation was developed between the SUDI index and the actual unplanned dilution (ELOS). The regression analysis showed that there is a relationship between SUDI and ELOS with a correlation coefficient (R2 = 0.90). With the use of this equation it is possible to determine ELOS according to the SUDI. Then, using the numerical simulation method with the 3DEC software, the dilution of stopes was simulated and the consequence of the effective parameters on the dilution was investigated, which resulted in providing a relationship for prediction of the dilution. Finally, using whit the datasets of 10 mentioned stopes the relationship was validated. The results show that there is a coefficient of determination of 0.82 between actual values and predicted values of dilution, which indicates that the predicted relationship is correct. Therefore, utilizing this equation it is possible to determine ELOS in relation to the effective parameters.