In this paper, the relationship between the rock mass rating system (RMR) and the productivity (U) of a double-shield tunnel boring machine has been investigated. For this purpose, first, daily excavation reports in the transfer tunnel and project delays in different formations of the Ab-e-Nosud RMR (southern part of Leileh) were examined and a 4.5 km length of this tunnel was recorded. The results showed that the overall productivity of the machine was 10% and the most delays were related to unfavorable geological conditions (GRRD). Next, the relationship between U and GRRD with RMR was evaluated using various forms of regression analysis. Among the empirical relationships developed, the best relationship for predicting U and GRRD in the studied tunnel is a quadratic polynomial function. Then, by creating a larger database including data from six tunneling projects, a comprehensive model was presented to evaluate the productivity of double-shield TBMs. This model is in logarithmic form and is able to provide an initial estimate of U for two-shield TBMs using RMR values. For this model, the statistical performance indicators, namely the coefficient of determination (R²), root mean square error (RMSE), and variance of error (VAF), were obtained to be 0.91, 0.003, and 0.91 percent, respectively. The results showed that all the equations proposed in this paper have acceptable accuracy and can provide an initial estimate of GRRD and U based on RMR values ​​with an acceptable error rate.