The mismatch between the number of crew and flight duties is widely occurred in airlines with multiple bases, which results in bad assignment at crew rostering stage. This problem can be fixed by generating partial pairing across bases, which is not considered in the traditional crew pairing generation method. Finding a globally optimal solution via the sequential approach may be impossible because the decision domain of the crew assignment problem was reduced by decisions made in the pairing problem. This paper proposes a new study of a crew pairing problem with partial pairing. A set partitioning model with partial pairing constraint based on segment time-space network firstly developed. To eliminate the effect on objective in subproblem caused by the additional constraint, a divide and conquer strategy was put forward. Finally, an empirical study was presented to validate the efficiency of the proposed algorithm. Compared with the traditional method, the method in this paper can significantly save costs and better meet the problem of crew mismatch.