Under the background of "flow space", secondary cities in urban network have attracted the attention from regional economic researchers. The role of intra-regional and inter-regional linkages in the development of secondary cities is related to the success or failure of inter-regional cooperation, but we have different understandings toward such a network. Based on the data of enterprises classified by machine learning, after industry classification, this paper studies the total changes and industry characteristics of urban linkages between Harbin and other cities in Northeast China. It is found that: (1) In the inward economic linkages of Harbin, the importance of cities in Northeast China is declining, and the hinterland of outward network is shrinking. (2) There is a trend of industrial-reversal and low-end in service industry in industrial linkages between Harbin and other cities in Northeast China. (3) Unlike other parts of the country, there is no "service center" in Northeast China. By studying the relationship between the total number of links and distance, gross regional product (GRP) and industry comparative advantages between Harbin and cities outside the northeast region, it is found that: (1) Harbin's linkage network shows scale-free characteristics, and the linkages with national central cities do not conform to the distance attenuation law; and the linkages are strongly linearly positively correlated with GRP. (2) The connection in advanced producer services shows a character of agglomeration, while the connection in manufacturing and other services presents a long tail distribution. Compared with the scope of contact, we should pay more attention to the quality of contact with a few "service centers".