A new synergy tracking method of infrared and radar is presented. To improve tracking accuracy, the unscented Kaiman filter (UKF) , which has better nonlinear approximation ability, is adopted. In addition, to reduce the possibility of radar being locked-on by adverse electronic support measure (ESM) , radar is under the intermittent-working state. After radar is turned off, the possible target position is estimated by a set of time polynomials, which is constructed based on the sufficient observations done before radar is turned off, the estimated values from time polynomials are compared with the current observation values from infrared to determine the time when radar is turned on. Simulation results show the method has a good tracking accuracy and effectively reduces the possibility of radar being locked-on by adverse ESM.