Focusing on the problem of uncertainty in the traceability of free-form surface contours measured using a surface structured light scanner， two different uncertainty evaluation models were proposed. The transparent box model， based on error source analysis， was used to calculate the matrix of transformation of the measurement point from the measurement coordinate system into the coordinate system in the CAD simulation. The measurement uncertainty was then analyzed according to the functional relationship of the surface profile， and the rotation and translation parameters obtained from the coordinate transformation. By contrast， the black box model， based on quantitative statistics， obtained the characteristic index of the measurement system from analysis of the measurement result. The evaluation model was established according to the indication， self-calibration， point cloud splicing， and point cloud registration errors， as well as the repeatability， reproducibility， and instrument resolution to enable uncertainty evaluation. The experimental results show that the uncertainty based on the transparent box model is 0.020 mm and the uncertainty based on the black box model is 0.023 mm. The quantitative statistical method is efficient and convenient for evaluating the uncertainty of measurements of a free surface obtained using the structured light scanning measurement system.