This research investigates the use of the Brainstorming method in the learning process to withdraw the best idea for improving students' Mathematical connection skills and the relation of Mathematical connection skills with metacognition skills. This study utilizes a quantitative method with a quasi-experimental model pretest-posttest control group design involving 60 Junior High School students in Sukabumi, West Java, Indonesia. The instrument used in this research is the developed mathematical connection skill test and the adapted metacognition questionnaire. Data analysis used to measure the comparison between control and experiment groups on students' Mathematical connection skills was the Mann-Whitney, and Correlation Product moment to analyze the correlation between Mathematical connection and metacognition on the experimental group. The result of the study shows that; (1) the Mathematical connection and metacognition skills of students who learn using the Brainstorming method is better than those who use expository learning. (2) there is an improvement in students' mathematical connection and metacognition skills by using the Brainstorming method (3) there is a relation between connection skills and the metacognition of students.