The application of Internet of Things (IoT) detection devices in the field of elderly healthy generates a large amount of detection data, which faces problems such as diversified data sources and non-uniform parsing methods. The fragmentation and confusion of the process from data subscription to data analysis have become important challenges affecting the development of the healthy aging field. In this paper, we designed and proposed an elderly health care framework with a model-as-a-service orientation. The framework automates the integration of health data for the elderly, including subscription services. It handles Model Training Analysis through a workflow-based system. Users can configure workflows using a low-code method provided by the platform. First, we abstracted and formalized the integration problem and workflow and defined a workflow description language. Then, we built an healthy aging unified modeling framework based on the problem model and provided a general introduction to the framework architecture. The module for analyzing is responsible for model training, persistence, and data analysis, enabling users to customize the selection of machine learning models and corresponding parametric. Ultimately, we built a model library and an external RESTful interface based on the automation option of Automated machine learning (AutoML). The experiments demonstrate this paper’s proposed framework’s feasibility and effectiveness.