To ensure the flight safety of unmanned aerial vehicles (UAV), the airborne intelligent condition monitoring has received a lot of attention. However, constrained by airborne computing resources, multi-task scenarios that monitor multiple key parameters pose a greater challenge to airborne real-time computing. To address this issue, this article proposes an airborne intelligent condition monitoring unit with multi-task scheduling optimization under limited field programmable gate array ( FPGA) resource. Firstly, the monitoring models of different scales corresponding to different monitoring parameters are established based on stacked long short-term memory, and the custom computing acceleration units for each model are constructed by using FPGA. Secondly, a multi-task scheduling optimization method with joint constraints of FPGA resources and model computing time is proposed to obtain customized computing acceleration unit deployment and computing scheduling strategies, which minimizes the completion time of all tasks. Finally, according to the above strategies, the acceleration units of specified scales are deployed in the airborne computing platform to complete multi-task scheduling and computing. Real flight data are used to verify the proposed method. The results show that the unit can efficiently perform real-time calculation for multiple condition monitoring tasks. The results show that the unit can efficiently perform real-time calculation for multiple condition monitoring tasks. Comparing with other FPGA-based computing methods, theefficiency is improved by 16. 08% .