In the traditional method, the measure of jobs-housing connection is to calculate the jobs-housing connection degree on the divided units, which lead to a certain subjectivity. As to the lack of relevant calculating methods, the different size of the units might lead to slight error. Therefore, we introduce Community Detection Algorithm, which is a kind of cluster analysis makes the intra-group connection largest and the inter-group connection smallest. According to the monthly work trip data of residents in Wuhan, we divide the jobs-housing connection units in central of Wuhan and analyze the effectiveness of the units. On this basis, we analyze the influence of natural and artificial boundaries on the division of jobs-housing connection units.