The primary mission of an emergency department (ED) is to treat the patients, to find out their diagnosis and to discharge them from system as possible as it can. The patients generally want to be informed the process to be applied and it can be done only with forming a systematic structure in ED. If the duration of the treatments is well-defined, the starting and finishing times of the processes can be known. Also, physicians try to reach the expected durations and they do not loose any unnecessary time. However, the patients who have the same complaints can react differently to the same treatments, and the procedure of the treatments varies from one patient to another, so it is not an easy problem to determine the durations. In this case Adaptive Neuro-Fuzzy Inference System (ANFIS) is used to determine the first care duration, which is the most important phase in an ED. Four effective factors are considered and used as inputs. MATLAB 7.0 fuzzy toolbox is used for the learning procedure. The results are compared with the original data. The results have shown that the predicted and the real values have a high correlation. © International Journal of Industrial Engineering.