In most of the commonly used approaches, the ranking of generalized fuzzy numbers is calculated based on the only one criterion i.e., rank criteria. However, there are some weaknesses associated with this criterion. In this paper, we discuss the insufficiency of considering only rank for the ordering of generalized fuzzy numbers. Some shortcomings in existing ranking methods have been pointed out. To overcome these drawbacks, we present a new ranking method which considers not only rank, but also some other criteria, such as mode, divergence and left or right spread of generalized fuzzy numbers, which are very important for the ordering of generalized fuzzy numbers. Several numerical examples are provided to illustrate the superiority of the proposed method. Some results in the support of the proposed method are also presented.