Given the drawbacks of the conventional Activity-Based Costing model (ABC), including the high costs of interviewing individuals, application of subjective and costly methods for the validation of allocations, as well as the maintenance and renovation issues, etc., the costing model researchers have proposed the Time-Driven Activity-Based Costing (TDABC) model or Time-Driven Intermediate Activity-Based Costing (TDIABC) model. In TDABC and TDIABC models, a linear association is presumed between cost and activity, while the cost function is not always linear. Artificial intelligence (AI) and fuzzy logic models could be used to resolve this issue and further improve the costing accuracy. Accordingly, the present study implemented the TDIABC method in combination with the Adaptive Neuro-Fuzzy Inference System (ANFIS) model for the first time and further embarked on a case study of the Hormozgan Province Gas Company. To this end, activity costs were determined based on the TDIABC, IABC, ABC, and TDIABC-ANFIS methods, respectively. At last, the costing results based on the Mean Absolute Deviation (MAD) were compared with the ABC method's results. The comparison results denoted that MAD in IABC is 0.74, 0.263 in TDIABC, and within the (0.245, 0.515) range in all structures designed via TDIABC-ANFIS. Hence, TDIABC-ANFIS costing is confirmed to be more accurate in comparison to the ABC, IABC, and TDIABC methods. Also, the Gauss-Transig-3 structure was determined to render the lowest error (0.245) among all structures designed with the TDIABC-ANFIS method.