The purpose of this research is to investigate how self-regulated learning affect group style and task style in the web-based learning environments. Theoretical framework for the study is based on web-based collaborative learning, cognition theory, task based learning and learner reflection. The importance of the learners' self-regulation in learning has been continuously reviewed by many educational researchers. In addition, it was supported by previous research results. Based on theoretical frameworks and previous researches on self-regulated learning provided design principles and strategies for cognitive tools supporting learner's self-regulative activities. As a result of this study, we have developed and applied to investigate (1) differences between the types of groups and the types of tasks on learner's task solving process, (2) differences between the types of groups and the types of tasks on learner's task solving performance. The two groups of learners regulated differently on concept comprehension and knowledge application task types. The total of 137 graduate students from A University participated in the study for the period of four weeks. Based on the type of self-regulation groups and tasks, participants were divided into four groups such as individual learning-concept comprehension, individual learning-knowledge application, collaborative learning-concept comprehension, and collaborative learning-knowledge application. The result of the study can be summarized as follows: First, in the study, the groups performing the same tasks did not show any differences in message frequencies. Second, for differences between the group performances, individual learners scored higher than collaborative groups in their achievement. For differences between the task types, however, knowledge application learners scored higher than those in concept comprehension. Although the differences were insignificant for individual and collaborative groups for concept comprehension tasks, however, knowledge application tasks showed significant differences in performances. The implication of the result indicated the effects of self-regulated learning in web-based learning seem significant for individual learning-knowledge application type of learning. In future research it is recommended that self-regulation supporting tool designs should consider components of group and task types.