To solve the inaccurate calculation problem caused by the large peak positioning error when the peak-to-peak (P2P) method is employed to demodulate low-finesse fiber-optic Fabry-Perot (FP) sensors, this paper proposes a P2P and interference-order positioning joint demodulation algorithm. For this purpose, two peaks are positioned in the reflection spectrum of the FP sensor, and cavity length is estimated by the conventional P2P method. Then, valley interference orders are introduced after a valley is positioned to generate a sequence of possible cavity length values corresponding to different interference orders. Finally, cavity length demodulation is achieved by retrieving the value in the cavity length sequence closest to the result estimated by the P2P method. To demonstrate the feasibility and superiority of the algorithm, this study also conducts demodulation simulations and experimental verifications of low-fineness FP sensors made from single-mode fibers. The experimental demodulation accuracy, better than 2.3 nm, is much higher than that of the conventional P2P method. The proposed algorithm can be used to accurately demodulate low-finesse FP sensors with a cavity length of 55-135 μm.