Objective To discuss the application effect of wearable devices in the management of home care services, and provide a more objective evaluation tool for evaluation of the quality of home care services. Methods Based on gender stratum, 36 nursing home nursing staffers were randomly divided into group A and group B. The two groups wore their devices and completed 9 nursing service items according to their own habits (day 1) and the post-training standards (day 2), respectively. Data sets A, B (untrained data) and data sets A′and B′(post-training data) were obtained. According to the experimental model of nursing behavior recognition, python program was used to complete the classification and recognition of data. Four groups of experiments were conducted, namely, experiment Train (A′), Test (A′-left) and Train (B′), Test (B′-left), experiment Train (A′), Test (B′), and Train (B′), Test (A′), experiment Train (A′,A,B), Test (B′), experiment Train (A′,B′,A,B), and Test (A-left) to obtain nursing behavior recognition accuracy (precision) values, f1_scores and recall values of nursing service behaviors, respectively. Results After training, the recognition accuracy (precision), f1_scores and recall values of the nine nursing service behaviors were all greater than 0.9600, and the highest could reach 1.000. In the worst case, the classification accuracy was at least 0.7522, and the use of standardized decomposition of nursing service action points through training could effectively improve the classification accuracy to more than 0.9821. Conclusion Wearable devices have a higher accuracy in identifying nursing behaviors and can be used for home care service management.