||Various learning strategies and methods were designed and implemented to improve students' learning performance. Among them, cooperative learning has been extensively researched for about twenty years. According to the research findings (Baloche, 1998; Kagan, 1994; Johnson & Johnson, 1993, 1994; Slavin, 1994), cooperative learning can be very effective in many ways if it is implemented in a proper way. In Cognitive domain, cooperative learning improves learners' academic performance. In psychomotor and affective domains, learners have more chances to learn different learning strategies from other team member(s), and have more chances to express their personal ideas and demonstrate their knowledge and skills. Through positive interdependence and interaction, most learners are more willing to participate the learning activities. Affectively, they feel less nervous and become happier in the process of learning. However, the success and effectiveness of cooperative learning depends on sound implementation of some essential factors, such as group goal, positive interdependence, interaction, grouping social skills, etc. Though cooperative learning has been widely applied in English classroom, most ESL teachers pay very Iittle attention to the effectiveness of grouping. A random grouping seemed to be the only way they used to group their students. However, random grouping is definitely not the only or the best way to do. We do believe that different criteria of grouping would make different influences on students' academic performance and attitude toward learning. How do most educators and scholars treat the effectiveness of grouping in cooperative learning? This paper aimed to provide a general and broad review of the issues of grouping in cooperative learning and discuss what are the alternatives that teachers should handle grouping issue in ESL classroom.