Abstract Implantable antenna is embedded in implantable medical devices to communicate with external world for continuous patient monitoring. The implantable antenna performance may vary with patients due to their different skin and fat thickness owing to their environment, location and concentration of collagen in skin, estrogen and testosterone concentration, age, diseases like obesity, Gottron syndrome etc. Estimation of IMA performance requires a proper description of its dependence on variable skin and fat thickness which is not studied extensively in previous related works. In this article, we have designed a muscle-implanted antenna at 2.45 GHz to examine its performance dependence on variable skin and fat thickness values. In this study, we have designed an artificial neural network (ANN) for prediction of antenna performance parameters at 2.45 GHz for wide range of skin and fat thickness variations. The designed ANN model can predict antenna performance with ∼0.99 % error. Here, accurate dependence analysis of radiation performance of implantable antenna in terms of S 11, realized gain, bandwidth and operating frequency on skin and fat thickness variations is performed which has not been studied yet and use of ANN in this field for accurate prediction is also a novel approach.