Fetal brain age prediction is crucial for assessing brain development and diagnosing congenital anomalies. Accurate gestational age estimation using imaging can enhance prenatal evaluation and understanding of brain maturity. The prediction of continuous values using deep learning remains a challenging task, despite the impressive application of convolutional neural networks (CNNs) for classification problems. The present study addresses this issue and uses transfer learning for such types of problem. The authors fine-Tuned ResNet50, DenseNet201, and MobileNetV2 by adding custom regression layers and selectively freezing pretrained layers to enhance training efficiency. Image resizing, normalization, and various data augmentation strategies were employed to avoid overfitting. Results show that fine-Tuning significantly improved regression accuracy, with further enhancement when the models were combined in an ensemble.