This systematic review investigates the application of deep learning in the classification and recognition of medicinal plant species. The study analyzed 31 primary studies published between January 2018 and December 2022. The findings reveal that 67.7% of the studies used private datasets, with 96.7% focusing on plant leaf organs, particularly leaf shapes for classification. Transfer learning with pre-trained models was used in 83.8% of the studies, and CNN was employed by 64.5% of the studies. The review highlights the need for globally available and publicly accessible medicinal plant datasets, as well as the trustworthiness of deep learning approaches for classification and recognition. The study also identifies challenges in dataset management, data preprocessing, and feature extraction, emphasizing the importance of interdisciplinary collaboration to address these issues. The results show that deep learning methods, such as CNN, have achieved high accuracy in classifying and recognizing medicinal plant species. However, challenges remain, including the need for high-quality labeled data, model interpretability, and standardized protocols for dataset preparation and preprocessing. The study concludes that deep learning offers significant potential for improving the classification and recognition of medicinal plant species, but further research is needed to address the identified challenges and enhance the reliability and effectiveness of deep learning approaches in this field.This systematic review investigates the application of deep learning in the classification and recognition of medicinal plant species. The study analyzed 31 primary studies published between January 2018 and December 2022. The findings reveal that 67.7% of the studies used private datasets, with 96.7% focusing on plant leaf organs, particularly leaf shapes for classification. Transfer learning with pre-trained models was used in 83.8% of the studies, and CNN was employed by 64.5% of the studies. The review highlights the need for globally available and publicly accessible medicinal plant datasets, as well as the trustworthiness of deep learning approaches for classification and recognition. The study also identifies challenges in dataset management, data preprocessing, and feature extraction, emphasizing the importance of interdisciplinary collaboration to address these issues. The results show that deep learning methods, such as CNN, have achieved high accuracy in classifying and recognizing medicinal plant species. However, challenges remain, including the need for high-quality labeled data, model interpretability, and standardized protocols for dataset preparation and preprocessing. The study concludes that deep learning offers significant potential for improving the classification and recognition of medicinal plant species, but further research is needed to address the identified challenges and enhance the reliability and effectiveness of deep learning approaches in this field.