Unlike most algorithms, multi-UAV system algorithms require a lot of time and effort to be validated to reach a conclusion about their robustness. The traditional way is to first run them in a virtual simulator and then put them on real UAVs to test them in practice. However, not only is this way time-consuming, but it is also dangerous since UAVs could be out of control when dealing with unexpected scenarios. In this paper, we propose a validation method for these algorithms using mixed reality framework. By design, the system algorithms are broken down into individual components leveraging heterogeneous drones. A general testbed is provided to verify their performances while injecting different types of faults in both simulation and the real world.