We present a model-constrained monoscopic (single-camera) phase-measuring deflectometry (PMD) framework to mitigate the height–slope ambiguity that fundamentally limits conventional single-view PMD. A priori surface knowledge, such as a nominal plane or sphere, is incorporated into a unified multi-factor iterative optimization that primarily refines the global height and pose parameters while compensating low-order reconstruction bias within a physically parameterized framework. The proposed method enables high-fidelity reconstruction without additional hardware, mechanical motion, or dual-camera calibration. Experiments on concave and planar mirrors validate the approach. For the concave mirror with a nominal radius of 151.5mm, the reconstructed radius deviates by less than 0.12% from the nominal value, with substantially improved form accuracy. Further analysis of the pose parameters demonstrates that accurate correction of tilt magnitude and azimuth is essential for suppressing low-order form errors in monoscopic PMD. In addition, repeated measurements verify the repeatability of the proposed framework. Compared with a binocular PMD baseline under the same fringe parameters, the proposed method achieves lower PV values and a fitted radius closer to the nominal value for the concave mirror while maintaining a compact single-camera configuration. This framework preserves the simplicity and flexibility of monoscopic PMD and provides a practical solution for high-precision, in-situ measurement of specular optical components.