Cyber-security is unavoidable in present days due to the security importance of each and every digital device part of a global network among millions of computing nodes. Industry 4.0 created much more impact on the automation of industrial micro functionalities to the manufacturing level, with more sensitive data to handle at every stage of organization, then security features of such platform are more important for a complete smooth working environment. This article will elaborate more on an understanding of Incremental learning with data analysis of huge Cyber-attack detection datasets which consist of Adware, Ransomware, SMS malware, scareware, Benign..etc. Incremental learning refers to a group of scalable various algorithms that as the ability to learn in terms of sequentially and can continuously update their models from huge infinite data streams. This working model helps to understand by adopting dynamic learning for varying incoming data inputs based on huge dataset inputs with scalable learning models of algorithms without discarding learned model outcomes. Number of incremental classification model were used where as ARFC gives very good results with higher accuracy, precision, recall, and f1-score values. The results revealed that the proposed model training generates a complete result by adopting variation in dynamic inputs which helps to improve classification higher accuracy by comparing with other training models. © 2022 IEEE.