In this study, capacity of multi-lane rotary intersections is discussed by using observations from four intersections in Izmir. Firstly, critical gap acceptance methods, especially gap forcing and reverse priority conditions are discussed. In the second part of the study, regression analysis methods, particularly TRL method, are introduced. A new capacity method which depends on regression analysis is also suggested. In the last part of the study, both methods are compared. As a result, both methods are found to be applicable but further research with increased number of observations is needed.