Abstract: Complex electromagnetic environment can easily interfere with unmanned aerial vehicle's (UAV's) datalink, which seriously threatens the flight safety of UAV. In this paper, a new method of UAV's datalink adaptive to electromagnetic interference (EMI) was proposed based on environmental perception. This method was arranged to improve the active anti-jamming ability of UAV. The corresponding EMI thresholds were made to change nonlinearly with the dynamic change of the UAV's operation signals. After selecting training sample variables and observation values, adopting a method of the relevance vector machine regression (RVR), a dynamic training model was established to reasonably predict the EMI thresholds under different operation signals. An expert system was established for UAV's datalink adapted for EMI. Then, the EMI characteristic parameters were determined by the airborne EMI environment monitoring platform. And the interfered levels of the datalink against EMI were distinguished based on the prediction of the target value by using the trained model. Combining the the technology schemes and the tactical schemes, the new method can be used to effectively eliminate the effect of the EMI on the UAV's datalink.