Background: Genetic neuromuscular diseases (NMD) are a heterogeneous group of disorders comprised hundreds of genes. Despite the advanced genetic testing modalities, about 40 % of patients with NMD do not have a diagnosis. Muscle MRI has been proven as a useful tool to orientate the genetic testing by looking at the muscle involvement severity pattern. Moreover, muscle MRI patterns can be specific and informative for muscular dystrophies and yet can be characteristic and diagnostic. Methods: Systematic review was conducted to review muscle MRI patters for all Limb Girdle Muscle Dystrophies (LGMD). Then, we applied artificial intelligence (AI) on muscle MRI patterns for LGMDs and other NMDs using open database containing muscle MRIs Mercuri scores from 950 individuals. Results: AI and machine learning were applied on 10 types of NMD muscle MRI Mercuri scores that represented muscle involvement based on the degree of fatty infiltration. Different models were generated, the one with highest accuracy was used. When tested on new patients, it achieved a 90% accuracy. Subsequently, was turned into a mobile application. Conclusions: Muscle MRI is a valuable tool to help in NMD diagnosis. Specific muscle involvement pattern can be predictive. Besides, AI facilitates the interpretation and comprehension of muscle imagining in NMD.