Distributed optical fiber Brillouin sensing has the unique ability of long-distance and high spatial resolution measurement, offering substantial potential across multiple fields. However, compared with point-based strain sensing techniques, the accuracy of high-resolution Brillouin strain measurements remains limited, due to challenges such as nonlinear scattering, low signal-to-noise ratio, and gain spectrum fitting error. Here we proposed and demonstrated the optically fused mechanism of Brillouin strain and Bragg grating strain in the same fiber, and developed a new transfer model for fidelity enhancement of distributed strain based on their high level of similitude. This transfer model combined affine transformation with adaptive weighting to effectively map strain instance vectors with large data gaps and significant dimension differences. For a 2.5-km precision fiber-optic coil used in underwater autonomous underwater vehicle (AUV) navigation, the average measurement error significantly decreased by 79% from 10.5% to 2.2% within 34.8 s compared to that before transfer. Only a small amount of reliable strain data was needed to improve the fidelity level of the distributed strain for each measurement. The sensing approach offers high accuracy, non-destructiveness, simple wiring, and low cost, making it suitable for long-distance underwater communication, geological disaster early warning, and aerospace structural fatigue monitoring.