Retinoblastoma is a type of eye cancer where cells in the retina begin to grow, or sometimes for some part(s) of eye the highly light-sensitive. The retina is made up of nerve tissue that senses light as it enters the front of the eye. This light is the main reason that stimulates the retina to send signals to the brain. The tumor usually begins to appear before the age of three, so early detection is essential by taking MRI scans and analyzing and comparing the images. In this article, an effective artificial intelligence system is built to detect retinoblastoma by taking MRI scans and then applying the CNN deep learning algorithm as well. There are two groups of images in the database, one of which includes images of abnormal eye tissues and cells and images of normal eye tissues and cells via magnetic resonance. All images used in this research are 225 × 225 pixels in JPEG format. The research aims to build a deep learning network capable of distinguishing patients with retinoblastoma, after training it to recognize this disease using deep learning algorithms, based on the latest developments in convolutional neural networks through the use of CNN. The discussion of the structure and work of different computer vision and some algorithms and techniques of deep learning and apply the convolutional neural network (CNN). The normalization is applied to the classification clusters based on DWT. As a result, the accuracy obtained is 99% and it is accepted, and it is better than the compared previous studies. This article was carried out using Simulink MATLAB R2023b.