Bone age assessment is a standardized clinical examination in pediatric endocrinology used for disease screening and growth development prediction. Traditional manual methods rely on physicians' experience in observing X-ray images of the left hand and wrist, which are inefficient and lack stability. While existing studies have proposed automated methods based on image processing or machine learning, their accuracy remains unsatisfactory. Inspired by the success of deep learning in image classification and speech recognition, this study develops a deep automated bone age assessment model based on convolutional neural networks (CNNs) and support vector regression (SVR) with multi-kernel learning (MKL). The model not only analyzes hand and wrist X-ray images but also integrates heterogeneous features such as race and gender. Experimental results demonstrate that this method achieves higher accuracy in bone age assessment compared to existing state-of-the-art techniques. For instance, a multi-scale multi-reception attention network (MMANet) proposed in 2023 enhances the recognition of key regional features and suppresses background interference, further improving the accuracy of bone age determination. However, the accuracy of AI-based bone age assessment varies significantly across different genders and ages, necessitating calibration. This approach offers a more objective and efficient method for bone age assessment, reducing clinicians' workload while improving efficiency and accuracy. However, the application of AI in bone age assessment is limited by the availability of high-quality, large-scale annotated image databases, posing challenges for widespread adoption.