A significant contribution to the medical healthcare has been made by machine learning (ML) and deep learning (DL). Both ML and DL benefited from the abundance of inclusive or unbiased big data sets. However, because there is less data available for rare diseases, ML and DL models became data deficient and were unable to detect such disorders. As a result, ML and DL failed to pay attention to rare disorders. The objective of this research work is to examine every possible route, from conventional learning to Meta-learning (MtL), for detecting rare ophthalmic disorders. This paper outlines the need of MtL framework’s usage to systematically examine and analyze rare ophthalmic disorders. To obtain the most relevant, reputable, and authentic literature, data sources such as IEEEXplore, Springer, Wiley, and Elsevier have been explored. The articles from the last 6 years have been taken to cover the most recent studies (from January 2017 to June 2022). The results of the research showed that because of the lack of data, DL neglected rare ophthalmological disorders. The proposed MtL-RODD framework employs a meta-learning approach to leverage information from multiple data sources, including clinical data, medical images, and electronic health records. The framework uses a deep neural network to extract features from these data sources and learns to diagnose rare ophthalmic disorders using a few-shot learning approach. An implementation of a proposed MtL-RODD framework is therefore suggested for the early diagnosis of rare disorders in ophthalmology in order to address these consequences.