Intrusion detection systems have been bolstering defences against cyber-attacks. VNF has emerged as a useful tool for delivering network operations across high-speed networks in cloud computing. Due to the shared nature of NFV deployment and the complexity of the underlying computer architecture, VNFs face greater security concerns than conventional network appliances. A universal intrusion detect mechanism is necessary for all Virtual Network Functions (VNFs), including virtual firewalls, routers, load balancers, and other devices. In this research, we present a model for intrusion detection in cloud-based VNF infrastructures using a lightweight machine learning approach. It is necessary to use a search technique to find the optimal subset of features inside the wrapper space, as this is what the suggested model relies on. Finally, practical characteristics are picked and used in conjunction with effective machine learning models for categorization. Python will be used to build the suggested model and then be compared to the accuracy, FAR, TPR, TNR, and FNR of several different current approaches are measured and compared.