Driven by the goals of carbon emission peak and carbon neutrality, the large-scale integration of power electronic-based renewable energy has significantly enhanced the dynamic response speed of the new power system, while concurrently increasing the risk of wideband oscillations. To effectively suppress the oscillations, there is an urgent need for fast, efficient, and precise detection of wideband oscillations. However, the existing detection algorithms struggle to simultaneously achieve rapidity, efficiency, and accuracy. Therefore, this paper proposes a fast detection method for wideband oscillations based on the hybrid sliding-window Fourier and frequency-locked algorithm, which incorporates the rapidity of the Fourier algorithm and the accuracy of the frequency-locked algorithm. Firstly, based on the instantaneous power theory, the voltage and current signals are converted into power signals, and the dual-frequency oscillation caused by frequency coupling is converted into a single-frequency oscillation. Secondly, the power signal is divided into low and high frequency bands to avoid mutual interference of signals during detection. For each frequency band, a single low-resolution sliding-window Fourier algorithm and a frequency-division sliding-window Fourier algorithm are employed to quickly obtain the approximate frequency and amplitude of the oscillation signal while significantly reducing the computational burden. Then, the approximate frequency is used as the center frequency of the phase-locked loop, and the oscillation amplitude is used as the trigger threshold of the phase-locked loop, so as to quickly lock the real frequency of the oscillation, and the detection time is stable within 10 ms. Finally, the effectiveness of the method is verified by MATLAB/Simulink electromagnetic transient simulation model and hardware-in-the-loop experiments.