Ear-EEG, as an emerging monitoring technology, has shown potential in the field of sleep monitoring due to its ease of use. This study aims to develop a complete portable Ear-EEG sleep monitoring system to achieve low-interference home sleep monitoring. The system designed and implemented in this study includes personalized Ear-EEG signal acquisition devices, data visualization modules, sleep staging models, and model deployment applications. The acquisition device obtains high-precision Ear-EEG signals in differential mode and transmits them via Bluetooth to the Upper Computer Software. The Upper Computer Software is responsible for data processing, visualization display, as well as the integration and application of the model. The sleep staging model combines ShuffleNet, Residual SE Block, and multi-head attention mechanism to improve model performance and classification accuracy. The trained model needs to be converted into ONNX format and deployed in the Upper Computer Software, ultimately achieving an integrated process from data collection to result presentation. Experimental results show that the developed Ear-EEG system can effectively monitor sleep and calculate sleep staging results while ensuring user comfort. The accuracy of the model in sleep stage classification is approximately 86%, with particularly outstanding performance in the recognition of wake, N2, N3, and REM (rapid eye movement) stages, although there is room for improvement in the classification of N1 stages. This system provides a new solution for home sleep monitoring, and the model demonstrates high performance after deployment, with broad application potential.