Diabetes is a prevalent chronic disease worldwide, which urgently requires effective blood glucose monitoring methods. This study proposes a multimodal non-invasive blood glucose monitoring system called Healthy-Ring, based on a smart ring. The key innovations of this system include: (1) the development of a conditional Wasserstein Generative Adversarial Network (PGAN) for high-fidelity reconstruction of PPG to ECG signals; (2) the introduction of a hybrid LSTM-Transformer architecture (GluFormer) that dynamically integrates ECG signals, capturing both blood glucose fluctuations and trends, with an accuracy of 82%; (3) the design of a low-power smart ring with integrated multimodal sensors. Experimental results show that the F1 score for detecting blood glucose abnormalities is 0.819, and the hardware cost is only 20% of that of traditional devices. This system provides a precise and convenient solution for diabetes management.