This study addresses critical challenges in teleoperation systems, focusing on dynamic delays and uncertainties that disrupt communication between master and slave robots. We propose advanced algorithms utilizing the Short-Time Fourier Transform and Taylor series for precise time-delay estimation, coupled with robust techniques for managing parametric uncertainties. By integrating Markov process models and non-linear filtering, our approach dynamically computes angular velocities for both master and slave robots, even in noisy environments. Simulation results demonstrate significant reductions in tracking errors, enhancing system stability and transparency. Experimental validation on a haptic device further confirms the real-world effectiveness of the proposed algorithms, showing superior performance compared to other control strategies under identical conditions. These contributions support the development of more resilient and efficient teleoperation systems for applications in hazardous or remote environments.