A dynamic model based on state space is presented for manufacturing network which performs as a kind of stochastic system. The model considers the influences of the internal and external disturbance factors on the network states,together with establishing network order transfer probability matrix from the order flow point of view and definition of the variables of manufacturing units in the network. Work-in-process is selected as the feedback variable of global optimal state from the system dynamic analysis point. Through the analysis of effects of the disturbance factors on order probability transfer matrix the state transfer equation is established and,the order transfer probability matrix is updated so as to simplify the solving complexity of state space of the manufacturing network. The order output is predicted with the proposed model using control theoretic methods,which provides distinction between with disturbances and without disturbances. Dynamic response of the disturbance factors of the manufacturing units is also analyzed with typical input signal. A CNC machining factory of airplane structure part is taken as an example to verify the effectiveness of the model and the perturbation analysis method.