To automatically sample mesoscale eddies by AUVs, the method to automatically recognize the dynamic features of eddies must be developed. In this study, a method for dynamic feature detection of mesoscale eddies was created based on SLA(Sea Level Anomaly)data. The main innovation is that a criterion to decide the succession relations of eddies was developed. With the succession relations of eddies being confirmed, we can calculate the area changing rate of eddy region, velocity of eddy centroid movement and some other dynamic features. Since the calculation cost of this algorithm is not enormous, it can be used in real time eddy tracking context, such as tracking eddies with AUVs. It is worth noting that SSH(Sea Surface Height)data can also be used in this algorithm.And SLA or SSH data can be acquired from the ocean numerical simulation model, satellite remote sensing or other methods. As long as SLA or SSH data are provided, our algorithm can be easily utilized for dynamic feature detection of mesoscale eddies.