The Ocean and Land Colour Instrument Imagery (OLCI) sensor provides moderate spatial and temporal resolution of marine data, becoming a promising tool for monitoring environmental changes in coastal waters. Therefore, it is fundamental to test and validate the resulting products from diverse algorithms to ensure the quality of the data. The complex waters of southern Buenos Aires Province inner and mid-shelf, characterized by the presence of estuaries and river inputs, are highly influenced by total suspended matter (TSM) variability. In this study, we evaluate the performance of three TSM products in different waters (estuarine, coastal and mid-shelf waters) with in situ data. Two products were obtained using neural networks (NN), i.e. OLCI L2 ESA standard product (TSM_NN) and Case 2 Regional Coast Colour processing (C2RCC_STD); and one product using the combination of an alternative Baseline Residual Atmospheric Correction approach and the Nechad 2010 TSM algorithm (BLR_NCHD). In general, TSM match-up results indicate that the OLCI TSM_NN and CR2CC_STD products are acceptable (R2 of 0.79-0.74, n=17, RMSE= 21-20 mg/L). The best results were obtained for BLR_NCHD product (R2=0.86, RMSE=7 mg/L). Future efforts needed to improve TSM retrieval involves the evaluation of the conversion factor between backscattering to TSM for the NN approaches and the evaluation of the atmospheric correction using in situ water reflectance measurements.