The satellite-Derived Bathymetry (SDB) has been researched for almost the past five decades, but the application of SDB for operational use is still in its initial stages. The higher concentration of suspended material is a major barrier in assessing bathymetry accurately with remote sensing data. The influence of various optically active substances (OAS) needs to be quantified to develop a more robust SDB algorithm. The goal of the study is to improve the accuracy of SDB by understanding the spatial distribution of Chlorophyll (Chl), Total Suspended Material (TSM), and Turbidity in coastal water. A preliminary investigation using the numerical analysis between Landsat-7 & 8 spectral bands, OAS parameters, and bathymetry is presented in this paper. SDB has been derived using three machine learning algorithms; Linear, Random Forest (RF), and Support Vector Machine (SVM) Regression. The resulting SDB residuals have been analyzed in relation to Chlorophyll, Turbidity, and TSM in coastal seawater. The study found that most of the erroneous SDB residuals and extreme values were distributed in high or medium concentrated OAS regions. The result of the study may further enhance SDB estimation accuracy by considering the influence of the above three OAS in coastal waters.