This paper aimed to identify the energy efficiency indices for rice producers in Golestan province, Iran. A non-linear Robust Data Envelopment Analysis (RDEA) model was applied based on seven energy inputs. The studied indices included energy efficiency, energy productivity, net energy efficiency, and water productivity. According to the results, all types of energy inputs and outputs were calculated 34,423.28 and 120,088.4 MJ/ha−1, respectively. Consumed nitrogen (29.9%), diesel (21.0%), and water for irrigation (20.5%) have the highest values of energy inputs. The results showed that at the level of conservatism Γ=7, half of the farmers are inefficient. The total optimum energy consumption was estimated to be 29,774 (MJ/ha−1), meaning that 10.01% of the input energy can be stored. Reducing energy waste can even improve the viability of rice farms and help them more control over energy consumption.