This paper studies the problem of plane search of unmanned aerial vehicles （UAVs） using dynamic sensors. A gradient-type nominal search strategy is studied. Since this strategy leads to local optimum which causes the suspension of the task， it does not guarantee to meet the goal of covering the search domain such that each point is surveyed for a certain level. A new strategy is consequently developed based on the nominal search strategy. Furthermore， a novel search strategy， taking advantage of area division， is proposed for the demand that the excessive search level and time cost are both as low as possible. Simulation results of a single UAV and multi-UAVs demonstrate the feasibility of the search strategies.