This dissertation proposes MPCC-DSVM (Model Predictive Current Control-Discrete Space Vector Modulation) and fault-tolerant operation which can secure the stability of the 3-leg 2-phase multi-load control systems, suitable for high-reliability mobility applications such as vehicles. The proposed MPCC-DSVM method eliminates the nonlinear modulation region and generates a virtual voltage vectors based on stationary reference voltage vector axis. Computational burden is reduced, enhancing dynamic response and calculation accuracy through cost function optimization by creating a candidate voltage vector in a specific region based on the estimated voltage angle using load parameters. In addition, a subdivided reference voltage is obtained by increasing the number of virtual voltage vectors through the proposed MPC-DSVM control method, the total harmonic distortion (THD) of the output current is reduced, which improves the power quality. The fault tolerant-operation detects current levels within 2 cycles in real-time, eliminating the need for complex coordinate transformation and equations. In the event of an open-circuit failure (OCF) in the 3-leg 2-phase inverter, the failure mode is defined within 10ms, and the duty ratios of the aphase, b-phase, and c-phase are adjusted accordingly. Simultaneously, MOSFET within the same leg are controlled through duty ratio modification and phase shift control to maintain system operation. Compared to Conventional-MPCC (C-MPCC), the proposed MPCC-DSVM method reduces the cost function, enhances calculation accuracy through virtual subdivided voltage vectors, and improves dynamic response and system stability by reducing THD. Fault- tolerant operation ensures the powernet reliability by maintaining the same voltage and current magnitude as pre-failure conditions in multiple load control. The proposed methods are verified through simulations and experimental results.