This report investigates the problem oflatency in a wireless network scenario, andproposes the idea of using Neural Networkmodels in a Digital Twin in order to predictthe latency in real-time.Different Digital Twin structures were proposed,including different amount of NeuralNetwork models as well as different inputs.To accompany the Digital Twin, a PhysicalTwin with a client-server file transmissionuse case was developed, in order to obtainvalues of latency. Based on data availablebefore a transmission, such as the physicallocation of the server and file size used, theDigital Twin was trained to predict.It was found that the Digital Twin wascapable of making predictions in undera millisecond by implementing the NeuralNetwork model as TensorFLow Lite models.This was significantly faster than PhysicalTwin in all scenarios, including when theobserved latency was the lowest. Thelatency prediction itself was successful, anda number of future considerations for moreaccurate predictions were proposed. Theseconsiderations include how to accommodatefor temporal characteristics in the observedlatency.