Considering the complexity and uncertainty of production planning in large manufacturing enterprise, hierarchical planning and scheduling is an effective solution to overcome the difficulty of scheduling both in long term and short term. Choosing a discrete task-oriented manufacturing factory as prototype, a hierarchical planning and scheduling model is developed, which consists of key resource leveling, detailed operation scheduling and workshop control. The upper level optimizes the bottleneck resource allocation based on aggregate networks of product, to make a feasible long term plan; short term schedule is then generated in medium level, using more detailed network; the schedule is finally executed in bottom level, which introduces a re-scheduling strategy to eliminate the plan error. To solve the model, the universal aggregate task network model and the algorithm for each level are researched. The simulation result demonstrated that the hierarchical method is valid and valuable in practice.