In this article the periodic signal detection method on the base of Duffing system chaotic oscillations analysis is presented. This work is a development of the chaos-based signal detection technique. Generally, chaos-based signal detection is the detection of chaotic-to-periodic state transition under input periodic component influence. If the input periodic component reaches certain threshold value, the system transforms from chaotic state to periodic state. The Duffing-type chaotic systems are often used for such a signal detection purpose because of their ability to work in chaotic state for a long time and relatively simple realization. The main advantage of chaos-based signal detection methods is the utilization of chaotic system sensitivity to weak signals. But such methods are not used in practice because of the chaotic system state control problems. The method presented does not require an exact system state control. The Duffing system works continuously in chaotic state and the periodic signal detection process is based on the analysis of Duffing system Poincare map fractal structure. This structure does not depend on noise, and therefore the minimum input signal-to-noise ratio required for periodic signal detection is not limited by chaotic system state control tolerance.