To address the issues of low sensitivity in high-velocity regions and limited upper measurement range in traditional fiber-optic thermal flow sensors, a high-sensitivity fiber-optic Fabry-Perot thermal flow velocity sensing method based on the Vernier effect and thermosensitive materials is proposed. The fiber-optic sensor employed in this method is fabricated by splicing a single-mode fiber (SMF) with a hollow-core fiber (HCF) filled with thermosensitive material at its end. The sensor achieves the first-stage sensitivity enhancement by utilizing the high temperature sensitivity of the thermosensitive material. Simultaneously, the cascaded Fabry-Perot interferometer (FPI) structure formed between the SMF and the thermosensitive material end-face creates the Vernier effect, which enables the second-stage sensitivity enhancement through the amplification characteristics of the Vernier effect. Through this two-stage enhancement mechanism, the sensor achieves improved flow velocity sensitivity in high-velocity regions and an extended measurement range. The sensing performance of the proposed high-sensitivity fiber-optic Fabry-Perot thermal flow velocity sensing method was theoretically analyzed using PDMS as the thermosensitive material. Concurrently, the sensor fabrication process was investigated, resulting in the successful fabrication of a physical sensor prototype. Experimental analysis was conducted to evaluate key sensing performance metrics, including sensitivity, maximum measurable flow velocity, and repeatability. The experimental results demonstrate that the sensor exhibits a temperature sensitivity of 1.399 nm/℃ and achieves a maximum measurable flow velocity of 25 m/s. Within the range of 17~25 m/s, the sensor′s response curve shows excellent linearity (R2=0.99), with a flow velocity sensitivity of 1.45 nm/(m/s). The repeatability deviation of sensitivity is only 1.24%, indicating excellent consistency. Owing to its high sensitivity, extended velocity measurement range, and compact form factor, this sensing method shows strong potential for industrial applications.