The energy of the AC grid is mainly stored in synchronous generators in the form of kinetic energy with a character of big inertia. In the study of AC grid stability, AC network is usually modelled as algebraic equations. While in DC grid, the energy is primarily stored in converter’s capacitor, as well as the DC network’s capacitor and inductor. The storage capacity of the DC network and the converter is even superior to the generator’s storage capacity in AC grid. Therefore, in the study of the DC grid stability, it is not proper to simplify the DC network as algebraic equation, but to take the dynamic differential equations of the DC network into consideration. In order to model any complex DC network precisely, a method using extended differential algebraic equations (DAE) to deduce the dynamic modelling of DC network is proposed. This method can eliminate the redundant state variables in the dynamic modelling of the DC network. Moreover, it can obtain the dynamic modelling of the DC network automatically through the submatrix junction and multiplication, thus save a lot of time in comparison to deduction of the state space system matrix manually. The accuracy of the method proposed is verified on a four-terminal DC network by comparing with the electromagnetic transient simulation and the small-signal simulation.