Abstract: In the case of ultra-low signal-noise-ratio (SNR), the conventional methods for chirp signal detection are invalid. An algorithm for weak chirp signal detection and parameter estimation is presented in this paper. The generalized filter array is constructed using a Duffing oscillator, and Lyapunov exponential is exploited to judge the state of the system. According to the characteristics of the chirp signal, the problem of chirp signal detection is transformed into a periodic signal detection problem to meet the demand of Duffing oscillator system. After the chirp rate is estimated and a new sequence is constructed by the corresponding processing, the starting frequency is estimated using the Duffing oscillator system again. Moreover, a segmented correlation smoothing method for enhancing the SNR is proposed to improve the performance of the method.