For a long time, the identification of open cluster member stars has been a challenge in the field of astronomy. Due to the complexity of the formation and evolution of open cluster, there is no unified method to accurately identify the member stars in open cluster. The objective is to select five dimensional parameters of the position and motion of stars from three different spatial distribution types of open star clusters. This study evaluates the performance of Density Based Spatial Clustering of Applications with Noise (DBSCAN), Friend-of-Friend (FOF) and Star's Galactic Origin (STAR GO) clustering methods in detecting open star clusters. The results show that the FOF and STAR GO algorithms are more suitable for clusters with special structure, and can identify the tidal or extended structure of the cluster, while DBSCAN can identify the member stars in the core region of the cluster more completely. The aim is to find a more balanced algorithm strategy between the details of the cluster structure and the integrity of the member star recognition.