This work introduces the need for the implementation of dynamic power network reduction. The current state of the art reveals the main interest is in three different reduction methods. Those are namely neural networks, system identification, and equivalent generator. By implementing specific algorithms, each procedure shows the capability of reproducing certain variables of the system. That enables dynamic reduction for the particular part of the system. The results show sufficient reduction with neural networks and system identification model. Still, the equivalent generator model shows somewhat worse results. Each method is trained and validated with specifically defined scenarios. The ending chapters present all of the results with valid comparisons in parallel.