In the medical field, Magnetic resonance image (MRI) brain tumor segmentation is highly crucial and important. Tumor segmentation can help in diagnosis and prognosis, overall growth predictions, Tumor density measures, and care treatment plans needed for patients. However, this task is extremely challenging due to low contrast, noise in medical images, and to the voluminous size of data. This work focuses on improving medical image semantic segmentation using a time-efficient ensemble learning approach. We propose a novel ensemble deep approach using a Convolutional Neural Network and three Autoencoders to extract relevant features from brain MRIs, reduce dimensionality, then apply supervised learning for pixel-by-pixel binary classification to achieve tumor segmentation. Experiments show promising results with our model achieving a Dice Similarity Coefficient of 88.63%, a Jaccard Similarity Index of 79.58% and a Recall of 84.34%, thus demonstrating the model’s ability to identify the tumor’s pixels with precision while consuming fewer resources. Therefore, the segmentation performance is sufficiently addressed in terms of accuracy, reliability and speed.