This thesis explores semi-active structural control methods for mitigating damage during seismic events. Semi-active devices offer the adaptability of active devices in conjunction with low power requirements and thus the reliability of passive devices. A number of structural applications utilising semi-active resetable devices in structural control are described and analysed. A distinguishing feature of this research is the novel design of a large-scale resetable device developed, manufactured and extensively tested. This design dramatically extends the capabilities of resetable devices by readily manipulating the device response to the structural demands and specific structural control requirements. In particular, the unique ability to use these devices to reshape or sculpt structural hysteretic behaviour offers significant new opportunities in semi-active structural control. The results indicate improvements in structural performance during seismic events is gained by approaches to structural control and enhanced damping methods that challenge conventional methods. Using an array of performance metrics the overall structural performance is examined without the typically narrow focus found in other studies. Suites of earthquake ground motion records are utilised to avoid bias to any particular type of motion and statistical analysis of the performance over these suites indicates the overall efficacy of the resetable devices in each case considered. A model that accurately captures all the device dynamics is developed, which can be used for a variety of device types and designs. In addition, the testing capabilities of structural control methods is enhanced by the development of a high speed, real-time hybrid test procedure providing a link between pure simulation and full-scale testing to increase confidence before investing in large experiments. Finally, the resetable devices are extended to improve the response force to size ratio, which additionally increases the force-displacement manipulation ability. Large-scale shake table experiments validate the findings of the analytical results. Very close correlation between analytical and experimental results including overall trends and numerical values verifies the analytical methods used and increases confidence in continuing research in this area. Furthermore, these large-scale experiments confirm the efficacy and accuracy of the the device model developed, leading to highly accurate quantitative prediction of the overall structural system response. Overall, this research presents a methodology for designing, testing and applying resetable devices in structural control. The devices developed in this research and the extensive modelling and testing dramatically extend the understanding and scope of these devices. Guidelines developed for these large-scale resetable device designs including a validated dynamic model brings the application of resetable devices closer to real structural control applications.