Because of their competitive advantages (e.g. cost and ease of installation), wireless smart sensors (WSS) have gained increasing attention for structural health monitoring applications. Various strategies have been developed to improve wireless data acquisition, in which measurement data is usually sent back in either centralized or decentralized way after data is completely acquired. However, inherent challenges remain for WSS to perform real-time data acquisition applications (e.g. real-time visualization of structural response). Although some efforts have been made to explore the strategies to address the challenges, obstacles still exist for high-throughput real-time data acquisition over large-scale network. Specifically, the event-driven operating systems (TinyOS) employed in most WSSs lack real-time scheduling support, imposing significant difficulties and limitations of real-time data acquisition. To address this obstacle, we consider the use of a real-time operating system, FreeRTOS, commonly used for industrial control systems, as an alternative solution. A framework of real-time data acquisition is presented and designed for high-throughput applications. In particular, preemptive multitasking was adopted to realize live streaming of a single sensor node, and Time Division Multiple Access was implemented to coordinate multiple sensor nodes within in a network. Lab tests were carried out to demonstrate that the developed framework can achieve high data-throughput for real-time applications.