The main purposes of this study are mainly three. They are firstly to explore the suitability of Main and Cassidy´s method of measurement of attachment of the six-year-old children, secondly to explore whether the responses of Chinese children could be classified into the secure, avoidance, ambivalence and control types according to the criteria developed by Main and Cassidy, and finally to study the relationship between the attachment and children´s social competencies in the school. The subjects of this study are 80 six-year-old children and their mothers. They are invited to the observation room located in the Department of Educational Psychology and Counseling, National Taiwan Normal University to participate in the measurement of attachment of the six-year-old children. Then the teachers of these subjects are asked to rate them with Preschool Social Competence Scale and Preschool Problem Behavior Check-list. Each subject is also given the Preschool Interpersonal Problem Solving Test and the Affective Role- Taking Test by the research assistants. The major findings are as the following: 1. The result has shown that there is a singnificant positive relationship between the security of attachment of the subjects and their reunion behavior toward mother, including looking at and smiling toward mother responding to mother´s initiating interaction with mother responding to mother´s invitation, and playing with mother. There is a significant negative relationship between the security of attachment of the subjects and their avoidance toward mother. 2. The subjects of this study could be classified into secure and insecure (avoidance) groups according to Main and Cassidy´ criteria. They are 65 (81.25%) and 15 (18.75%) subjects in each group. However, the ambivalence and control groups have not been found. 3. The subjects in the secure group are further divided into very high secure, highly secure, median secure and low secure subgroups. There are five groups of subjects together with the insecure group. Their social competence scores have been compared. The result has shown that there are no signifigent differences among groups regarding to the social competence, problem behavior and affective role-taking ability. However, the subjects of the very high secure group did show higher scores on the interpersonal problem solving test than the insecure group. They give more prosocial strategies. They have shown higher score on fluency and flexibility.