The popularity of automatic modulation categorization (AMC) is high in recent years owing to the many advantages. When it comes to communication, reliability in an AMC is very critical. Increasing the number of signals exponentially increases the cost of using the AMC. Precise classification methods, such as neural networks, in which either the parameters of the neural network or the dimensions of the input or output variables are modified dynamically, are not successful in obtaining high accuracy results. To improve the accuracy of the modulation categorization, this study employs a "QIGA" feature selection model based on a Quantum (inspired) Genetic Algorithm (QIGA). QIGA is used to choose the correct functionality and to limit the number of examples that must be learned so that the overall system time is shortened and the cost of computing is reduced. Selecting excellent characteristics is enhanced via quantum computing and this is done to lower the complexity of the solutions. The internal validation results demonstrated that the QIGA model significantly improved the statistical match quality and significantly outperformed the other models.