Purpose/Significance To explore deep learning models can identify care needs entities in user needs texts in home healthcare scenarios, so as to accurately identify user needs through automated means to provide technical support for improving the efficiency and quality of home care services. Method/Process 560 pieces of objective data of user care needs are selected to classify and label the care needs entities in the texts based on the International Classification of Functioning, Disability and Health. The BERT-BiLSTM-CRF model is used to perform the entity recognition task. The effect of the model is verified by the ablation experiment, the experimental results are analyzed, and the performance of the model is evaluated. Result/Conclusion The model’s precision, recall, and F1-score at the entity-level micro-average are 0.752 9,0.775 8, and 0.764 2, respectively, indicating that the model can provide strong support for automatic mining of user needs, optimization of care service process and improvement of care quality in home healthcare scenarios.