Underwater acoustic communication (UWA) technologies are critical for advancing ocean research and exploring aquatic environments. These technologies face significant challenges due to environmental factors such as salinity, temperature, pressure, and noise levels. The underwater acoustic channel (UWAC) is particularly complex, characterized by time-varying properties and high multipath spread, making channel modeling and estimation a challenging task. Accurate channel state information (CSI) is essential for effective receiver design and channel capacity analysis in underwater communication systems. This thesis presents a comprehensive review of the latest advancements in channel estimation techniques for UWA communication, systematically evaluating and contrasting earlier works. It provides an in-depth overview of various channel models, techniques, and algorithms for channel equalization and estimation. Specifically, chapter three introduces a low-complexity detection approach for Internet of Underwater Things (IoUT) connectivity. This approach employs single carrier frequency domain equalization (FDE) combined with amplify-and-forward protocols to streamline computational operations at the transmitter and sensor nodes. The use of fast Fourier transform (FFT) and cyclic prefix (CP)is proposed to simplify the equalization process, and adaptive frequency domain equalization (FDE) is suggested to address channels with rapid Doppler shifts. Two adaptive detectors based on recursive least squares (RLS) and least mean squares (LMS) algorithms are proposed, with numerical simulations demonstrating their near-optimal bit error rate performance. The thesis also tackles the challenge of accurately representing the non-uniform sparse nature of the UWAC, which increases algorithm complexity and computation time. A two-dimensional frequency domain approach is proposed, incorporating both primary and auxiliary channels to enhance channel estimation while reducing computational demands. Chapter four showcases the efficacy of this method, achieving superior mean square error(MSE) and execution time compared to traditional methods. Furthermore, the thesis develops a highly flexible prototype for underwater channel communication, integrating deep learning (DL) techniques based on long short-term memory(LSTM) systems for UWAC estimation. Using Simulink software tools, the LSTM algorithm is employed to estimate underwater channel characteristics in an end-to-end fashion, encompassing both training and testing datasets. This approach facilitates narrowband communications and simplifies coherent signal detection by leveraging the coherence be-tween consecutive systems, thereby eliminating the need for direct estimation of under-water channel coefficients. Finally, the thesis concludes with a summary of findings and suggestions for future research directions. Simulation results throughout the work high-light the effectiveness of the proposed methods, particularly in comparison to LS and MMSE algorithms.