Software Defined Networks (SDN) have provided an advancement in terms of flexibility and scalability in recent times in networking. SDN paves the way for a centralized monitoring system. This centralisation has also led to increased security threats which has made the network monitors to focus on the improved security of the systems with utmost priority. The proposed methodology simulates an SDN environment using two Virtual Machines with Mininet and RYU controller. A traffic is generated to create a false attack and the detection is visualized using this environment. Machine Learning techniques provide an efficient and logical way to implement a solution. Feature Selection is an important aspect and Recursive Feature Elimination with Cross-Validation (RFECV) is used to get the final subset of features based on the importance of their significance. Synthetic Minority Oversampling Technique (SMOTE) is used in data preprocessing to handle the data imbalance. Models like XGBoost, Random Forest and Convolutional Neural Networks are used. Using ensemble technique, a meta-learner is constructed using Random Forest, XGBoost and AdaBoost to use the strengths of the standalone models and create the best version of the classifier that shows 98.8% accuracy. Accuracy, F1 score, Precision and Recall are used to evaluate and compare the performance of the models. The generalization of the constructed meta-learner is tested on another dataset and the results show good accuracy. This model can be integrated into real-world networks to enhance security by accurate intrusion detection.