Brain Computer Interface (BCI) is the fast-growing research area that focuses on establishing the symbiotic relation between human brain and artificial intelligence. BCI enable the users to directly communicate with a computer by means of brain activity. Disability of a person due to medical issues such as spinal cord injury or cervical issues makes him/her to depend on caretakers for operating the home appliances. With the advancements in BCI technology, disabled people can control the home appliances through their brain activity without depending on caretaker. The existing models have accessibility challenges for the disabled people in two aspects Viz. accuracy and interaction time. The lack of high prediction accuracy as well as longer interaction time makes the existing models ineffective for the usage. To address the issues, hybridized deep learning model comprising Stacked Denoising Autoencoders and Extreme Machine learning is proposed and examined in this research. The proposed model acquires and preprocess the brain data of user and predicts the user intention with the improved accuracy. The proposed model is compared with different existing models investigated in the literature. The proposed approach achieves the maximum testing accuracy of 91.8% with 70% of training (10,500 Sample) and 30% of test validation (4,500 Sample) on the brain data of single user and takes interaction time of 0.48 seconds. The evaluation results of model validate the effectiveness of the proposed methodology.