Abstract : Deep learning is a branch of artificial intelligence. Image processing has been revolutionized with the advent of deep learning. It is making its big way in data analysis in medicine. With the rise of deep neural networks in deep learning, the accuracy of predicting a task has improved in a tremendous manner by incorporating deep layers into the model that allow the system to learn complex data. Usage of the CNN in daily practice warrants its validation done external data set before widespread application. A systematic literature review on deep learning for fracture diagnosis was undertaken on the basis of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. English articles with publication dates between 2000 and 2022 were evaluated. Inclusion criteria were studies of deep learning applied to fracture diagnosis with (a) a similar group, encompassing a comparable research purpose and cohort, and (b) sufficient and calculable data to analyze outcomes. There were 903 publications found after a literature search. Seven studies that are included in the syetmatic review were published between 2020 and 2022. In conclusion Deep learning has shown considerable potential and feasibility in fracture screening, emergency department diagnosis, workflow prioritisation, mistake minimization, and speciality physician training despite existing constraints. Currently, the diagnostic precision is comparable to that of general practitioners. There is need for sound external validation before acceptance of the CNN as independent diagnostic modality.