In recent years, deep learning models, such as the UNet architecture, have shown great potential in medical image segmentation tasks. However, these models often require large amounts of annotated data to achieve high performance. In this paper, we propose a novel approach that combines the UNet architecture with the Vision Transformer with Multi-Head Attention model (MAViT-UNet) for medical image segmentation. Our method leverages the ability of ViT to capture long-range dependencies in images, while still benefitting from the UNet's skip connections for local feature extraction. Through extensive experiments on various medical datasets, we demonstrate that our proposed Vision Transformer based UNet achieves state-of-the-art performance in medical image segmentation tasks, even with limited annotated data. The results suggest the potential of transformer-based models in the field of medical image analysis.