As the sustainable development movement gains traction, many initiatives have aimed to provide support with sustainability assessment methods using a loosely structured set of indicators, resulting in an explosive development of such indicators. Several reviews have sought to gain a clearer understanding of this “indicator zoo” by comparing assessment methods. However, most reviews covered few methods and mainly focused on describing and evaluating them to help users select the one that suited their needs. What is still lacking is an in-depth analysis of the conceptual and methodological framework behind the sustainability assessment methods in order to provide recommendations for assessment method developers and identify research gaps. To fill this gap, we conducted a conceptual and methodological review on a sample of 262 studies covering the worldwide agricultural sector between 1993 and 2019. Because the subject is so vast and due to disciplinary barriers, we restricted this review to studies with an environmental dimension but did not set any geographic limitations (the studies cover both temperate and tropical zones). The initial results show a need for clarity on terminology. Methods should also explicitly specify the purpose, target users and temporal scale. Additional efforts must be made in selecting indicators on the causal chain at emission level or those that pertain to impacts, such as in life cycle assessment. There are additional research gaps when more than a simple sum of scores is required to aggregate indicators, or when dealing with upscaling or spatialization issues. Dynamic assessment of resilience or robustness with multiple criteria also remains a challenge. Environmental sustainability assessment methods will also need to integrate ecosystem services and emerging environmental issues such as pollution from antibiotics or microplastics. Finally, a similar study on assessment methods should be conducted with a focus on social and economic dimensions.