High-aspect-ratio microholes are widely used in aerospace, electronic packaging, biomedical devices, and precision manufacturing. However, their non-contact optical measurement is still challenged by weak bottom signals, false-focus interference caused by stray reflections and scattering inside the hole, and accumulated uncertainty in aspect-ratio evaluation. To address these problems, this paper proposes a multi-feature fusion optical measurement method based on an aspect-ratio constraint for stable measurement of microhole depth and aspect ratio. In this method, a sequence of microscopic images is acquired at different defocus positions through precise axial scanning. Within a predefined region of interest, multiple image features, including sharpness, mean gray level, and contrast, are extracted to establish the response relationship between image characteristics and axial displacement. After min–max normalization, these features are adaptively weighted according to their response dispersion and fused into a comprehensive feature response function. The local extrema of the fused response curve are then extracted as candidate bottom positions. Furthermore, combined with the measured entrance diameter, a reasonable prior interval of the aspect ratio is introduced as a geometric constraint to screen the candidate depths, thereby reducing the uncertainty caused by false-focus points and single-feature judgment. Experiments were carried out on 20 microhole samples with different aspect ratios. The results show that the proposed method can effectively suppress false-focus interference under weak illumination, low contrast, and complex internal reflection conditions. A good linear agreement is obtained between the measured aspect ratios and the standard values, with a fitted slope of 0.997 and an intercept of 0.00912. The mean absolute error, root-mean-square error, and maximum absolute error are 0.0195, 0.0235, and 0.06, respectively. The proposed method provides a feasible and robust approach for non-contact measurement of geometric parameters of high-aspect-ratio microholes and has potential applications in online inspection, quality evaluation, and process optimization in precision manufacturing.