Modern medical devices have the ability to producing and transmitting all kinds of basic medical data. These devices can share data with each other or send data to a central platform in the cloud. The new trend in the healthcare industry is the creation of a medical monitoring system integrated with electronic medical records that presented from the cloud, that can be accessed anytime and anywhere. It is essential to store, transfer, share and visualize large volume heterogeneous medical data in a low-cost, fast and secure database system. In this study, a medical Internet of Things (mIoT) platform that can use different database systems was realized in order to acquisition heterogeneous medical data from sensors to visualize and store data. Four different database models on mIoT platform were tested with four different scenarios. The performance of the database models used in the mIoT platform in these scenarios; query time, data preparation, flexibility, security and scalability parameters were compared. It was observed that the non-relational database model (NoSQL: Not only Structured Query Language) used on the mIoT platform was more efficient in reading/writing operations than relational database models, and its performance, flexibility and scalability were better than relational database systems.