Measuring biophysical indicators of carbon storage by forest vegetation, e.g., tree canopy height, growing stock volume, and aboveground biomass (AGB) is useful in understanding the carbon balance in terrestrial environment. Estimating the carbon content of forests accurately provides critical inputs for United Nations Framework Convention on Climate Change (UNFCCC), as this carbon content estimate forms the basis of reporting for initiatives developed under the UNFCCC, such as Clean Development Mechanisms (CDM) and the Reducing Emissions from Deforestation and forest Degradation (REDD) in developing countries. Forest biomass can be estimated in mainly two ways, field-based measurements and remotely sensed methods. Despite its exceptional accuracy, it is not always feasible to measure biomass based on field-based methods only. Conducting field measurements over a regional scale demands significant time and resources and is often impractical in inaccessible locations. Remote sensing, however, offers an efficient approach for assessing and monitoring biomass at broader spatial scales. This study presents how accurately the AGB of a mangrove forest can be estimated, using C band SAR backscatter alongside with textural parameters and optical data-derived vegetation indices. This study demonstrated that machine learning methods can establish a complex non-linear relationship between the predictor and predicted variables and thereby offers improved accuracy in AGB estimates.