The detection of weak harmonic signals has important practical value. In this paper, the detection of weak harmonic signals in strong noise is realized with the controlled Chen's system. Dynamics can be divided into slowly varying dynamics and fast varying dynamics, so a slowly varying dynamics is obtained by an averaging method. The critical values of detection parameters are determined, which lead to a sudden change of system dynamical behavior from periodic orbit to stable equilibrium point. Weak harmonic signals can be detected by adjusting the detection parameters and observing the change of system variables. Simulation results show that weak harmonic signals in strong noise can be detected accurately with this system. Compared with existing detection methods with chaotic oscillator, this method is of stronger immunity to noise, and the accurate range of parameter threshold can be obtained through theoretical analysis, which enables its popularization and application in relevant fields.