Tumor is the deadliest disease of today’s time. It can occur in any part of the body. Brain tumor is one such type of tumor a body can have. It can be benign or malignant; benign tumor is not lethal but malignant tumor can cause risk to human life. It can happen in the nervous system of the human body as well as in the cortex. The growth of tumors in our body is so fast that it becomes crucial to detect and treat it early. The involvement of intelligent methods made it possible to diagnose such types of diseases on time. The evolution of machine learning (ML), a subfield of artificial intelligence (AI), made it possible to diagnose any disease at an early stage and treat it effectively, especially in case of deadly diseases such as tumors. To detect the tumor, we generally need MRI images or CT scan images. Because analyzing MRI is faster and more efficient. Deep learning is very efficient in image classification. Many deep learning models are present today to diagnose it. CNN is one such deep learning model and it performs exceptionally well in analyzing biological images, but there are some disadvantages related to CNN. One such disadvantage of CNN is that it fails to encode the position and orientation of objects. Capsule network provides a solution to the disadvantage of CNN. A capsule network is a successor of CNN that can closely trace biological neural networks and we can use it for biological image classification. It is not affected by image orientation. The paper presents a systematic review of the role of the capsule network in detecting brain tumor.