This paper proposes a novel deep learning framework named DL-GCN for indoor temperature prediction in multi-zone buildings. The framework combines distributed Long Short-Term Memory (LSTM) and Graph Convolutional Networks (GCN) to capture spatial-temporal correlations and multivariable coupling features. The GCN is used to extract spatial features from temperature and humidity data, while the distributed LSTM module fuses other data like light and AC power consumption to learn temporal and coupling interactions. The model was validated using real datasets from a large-scale building, showing superior performance in multi-zone indoor temperature prediction compared to baseline models. The study highlights the importance of spatial and temporal features in accurate temperature forecasting and demonstrates the effectiveness of the DL-GCN architecture in handling complex spatial-temporal correlations and multivariable interactions. The results confirm that the proposed model outperforms existing methods in terms of prediction accuracy, making it a promising solution for optimizing HVAC systems in large buildings.This paper proposes a novel deep learning framework named DL-GCN for indoor temperature prediction in multi-zone buildings. The framework combines distributed Long Short-Term Memory (LSTM) and Graph Convolutional Networks (GCN) to capture spatial-temporal correlations and multivariable coupling features. The GCN is used to extract spatial features from temperature and humidity data, while the distributed LSTM module fuses other data like light and AC power consumption to learn temporal and coupling interactions. The model was validated using real datasets from a large-scale building, showing superior performance in multi-zone indoor temperature prediction compared to baseline models. The study highlights the importance of spatial and temporal features in accurate temperature forecasting and demonstrates the effectiveness of the DL-GCN architecture in handling complex spatial-temporal correlations and multivariable interactions. The results confirm that the proposed model outperforms existing methods in terms of prediction accuracy, making it a promising solution for optimizing HVAC systems in large buildings.