To handle uncertain terms, disturbance and time-delay in a finite-time sense simultaneously, an adaptive barrier sliding-mode control (SMC) combined with a state predictor is designed for delayed dynamics with unknown disturbances and parametric variations. Traditional SMCs often suffer from time-delay and require prior knowledge of disturbance bounds. To address these limitations, a state predictor is integrated to estimate system uncertainties in real-time, reducing reliance on conservative assumptions. This approach does not require a priori knowledge of disturbance magnitudes and actively compensates for time-delay without exact system model. Additionally, an adaptive barrier function is introduced to adjust the sliding surface dynamically, ensuring finite-time convergence while preventing excessive control effort. The proposed control strategy guarantees robustness against uncertainties, with rigorous stability analysis confirming the ultimate boundedness of the overall system. Simulation and comparative studies manifest the efficacy of the introduced controller in improving tracking accuracy and enhancing adaptability in the presence of disturbances. The results highlight its potential for application in typical control systems.