A significant challenge in deploying bipedal robots for human-oriented real-world applications is their ability to maintain balance when externally disturbed. Current momentum-based balance control strategies often exhibit inadequate robustness to disturbances due to reliance on simple proportional controllers and imprecise incorporation of desired angular momentum changes. Furthermore, the sequential activation of momentum and posture correction controllers compromises system stability when confronted with consecutive disturbances. This paper proposes and validates a new Variable Power Reaching Law for Sliding Mode Control (SMC) to enhance the regulation of linear momentum against disturbances. The proposed reaching law adjusts dynamically to the system's errors, ensuring fast convergence and minimal chattering. In this paper, we precisely define the desired angular momentum change in relation to the Center of Pressure (CoP), a crucial stability metric, as well as the desired linear momentum and ground reaction forces. The null-space method, which allows for simultaneous task execution by using unused degrees of freedom, is employed to ensure effective balance and upright posture without interference. The posture correction control is projected onto the null-space of momentum control. Simulation results confirm that the proposed control system effectively stabilizes the robot against external disturbances, regulating momentum and restoring upright posture. The null-space method proves effective in maintaining balance under multiple disturbances by simultaneously controlling momentum and posture. Comparative evaluations show that our approach outperforms traditional momentum-based controls and nonadaptive reaching laws, reducing CoP fluctuations, managing disturbances up to 117 N, and minimizing chattering and steady-state error. These advancements underscore the potential for deploying bipedal robots in dynamic environments.