The rapid expansion of wireless technology has caused a significant increase in the number of end-users, leading to a higher level of congestion in the spectrum. Ensuring a reliable and consistent service has become an increasingly difficult challenge. There is an urgent requirement for intelligent communication systems in order to improve the quality of service. Automatic modulation recognition is an essential problem in advanced wireless technologies like cognitive radio and software-defined radios. Cognitive radio enables dynamic modulation and demodulation, enabling it to see, acquire knowledge from, and adjust to its surroundings. Automatic modulation recognition can be simplified to a classification problem. Recent advancements in deep learning have showcased exceptional ability in a wide range of classification problems. This study introduces a modulation classification model based on deep neural networks. The approach aims to automatically identify ten different forms of radio signal modulations, including both analog and digital modes. The novel neural network architecture we have developed synergistically integrates the capabilities of a Convolutional Neural Network and long short-term memory to accurately categorize modulation types of radio signals. We conducted thorough trials on the RadioML2016.10b dataset, which resulted in improved performance in modulation recognition when compared to other models.