This study describes the results of combining two Modified Packed Ucell (MPUC) inverters to construct a cascaded MPUC network. With only 12 switching combinations, this topology can create up to 49 levels of voltages. The voltages of two MPUCs are adjusted such that each auxiliary source has one-third of the voltage level of the primary source and the primary source voltage level of the MPUC is one-seventh of the value of the DC voltage of the cascaded unit. The formation of different voltage levels by cascading is presented in detailed manner. An Artificial Neural Network (ANN) trained for harmonic removal in the output voltage waveform is used to control this converter. ANN is trained on two different frequency switching schemes. The cascaded unit with high voltage stress (14VDC) is switched at low frequency, low switching frequency operation of the MPUC, results in decreased power dissipation and higher system efficiency, and the multilayer output voltage waveform is virtually sinusoidal with negligible Total Harmonic Distortion (THD) content (less than 2%). The presented inverter and its ANN based control method result in a design trade-off between filter size and switching frequency (efficiency). This is predominantly helpful for high power applications, such as HVDC transmission system. The simulation realized for converter and its control in MATLAB®/Simulink environment show good agreement with theoretical values.