Fifth Generation (5G) networks provide data communications through various latest technologies including Software Defined Network (SDN), Artificial Intelligence, Machine Learning and Cloud Computing. In 5G, secure data communication is a challenging issue due to the presence of enormous volume of users including malicious users communicating with latest technologies and also based their own requirements. In such a scenario, fuzzy rules and cryptographic techniques can play a major role in providing security to the data which are either communicated through the network or stored in network based databases including distributed databases and cloud databases with cloud networks. Therefore, new and efficient mechanisms for generation and exchange of keys are necessary since they are the most important component of cryptographic methods. Since most of the existing key generation techniques are focusing on 3Gand 4G networks, new key generation methods that can be generalized to n-th order polynomials are necessary to suit the security requirements of 5G networks which is smart by using rules from Artificial Intelligence. This paper proposes a new key generation and encryption/decryption mechanism which is based on both symmetric key cryptography and polynomial operations for providing effective security on data communication in 5G networks. In this work, we introduce the usage of fuzzy rules and Binomial Theorem (Pascal triangle) technique for performing the data encryption process more efficiently since it is not used in any of the existing cryptographic algorithms. Moreover, two different polynomial equations, one of degree three and another of degree two are used in the proposed work for effective key generation. Here, we have applied differential calculus for finding the second-degree polynomial. In the decryption part of the proposed mechanism, n th root operation is applied which is able to reduce the number of steps used in a single mode operation. The experimental results of the proposed work proved that the proposed security model with fuzzy rule-based approach is better than other related systems that are available in the literature in terms of reduction in computational complexity and increase in security.