The pull-push hybrid P2P streaming, as an emerging and promising approach, has achieved some success in delivering live video traffic. The sub-stream scheduling problem is a key design issue in a hybrid system. In this paper, we propose a max-flow model for unifying this sub-stream scheduling problem. We find that the sub-stream scheduling problem in GridMedia, CoolStreaming+ and LStreaming can be formulated into a special case of the proposed max-flow model. We further propose a min-cost flow model to combat peer heterogeneity in scheduling sub-streams. This min-cost flow model is implemented in a prototype system, LStreaming+. The accuracy of the max-flow model and the outstanding performance of LStreaming+ are demonstrated by extensive simulations. We also show that LStreaming+ achieves excellent performance in prototype experiments.