As processing cost became more affordable, the telecommunications market started to replace one-purpose hardwares, e.g., modulators and amplifiers, by their software implementation version creating Software Defined Radio (SDR) systems. Some of the advantages of SDR systems are the low cost to develop new functionalities and the short time to market by quickly making new software available to be downloaded. In this context, the present work used an SDR platform to deploy a Long Term Evolution (LTE) network testbed and evaluate an algorithm to predict the signal quality of a link between an User Equipment (UE) and an Evolved Node B (eNB). A performance evaluation showed that low error of prediction can be achieved based on the choice of the algorithm parameters, which can be configured according to the environment conditions.