Growth in technology has accelerated in the last few decades, and therefore demand for energy has dramatically increased. This has also brought about a major problem called global warming. The scarcity of resources, indeed, have been compelling us to sustain energy with the adoption of cleaner energy practices and alternative ways that can help us reduce energy consumption and carbon dioxide emissions. With the pressure due to increasing global environmental consciousness, the international shipping industry has thus been exposed to some challenges and energy-related problems it needs to wisely deal with. To reduce fuel consumption and greenhouse gases, ship owners and operators need to find an efficient fuel economy with the control and alteration of some related factors. In this study, several determinants of fuel consumption in maritime transport were considered and alternative models were built employing the Response Surface Methodology using data obtained from the noon reports of a bulk carrier. Since the model building process is relatively tedious in the Response Surface Methodology, a guideline was also provided. The results indicate that as their speed increases up to a certain point, ships consume less fuel; however, after such a threshold, fuel consumption begins to increase. In addition, even though its main effect was found insignificant, revolution per minute's interaction with other inputs was found to affect the output. Thus, adopting the Response Surface Methodology, which allows for quadratic terms and higher order interactions, is critical to build a predictive model that performs well so that ship operators and shipping companies can make economic and environment-friendly decisions better. This study serves the literature by estimating the fuel consumptions of cargo ships using statistical models so that carbon dioxide emissions from fuel consumption can be estimated. Even though specific types of fuel and ships were considered, the approach adopted here can be extended and applied to other cases as long as there is accurate data. (C) 2020 Elsevier Ltd. All rights reserved.