Laser light reflection during the laser transmission welding (LTW) of thermoplastics has the potential to overheat and/or cause unintentional welding of adjacent features of the part being welded. For this reason, and in order to assess how much light is being absorbed by the transparent part (after measurement of the light transmitted through the transparent part), it is important to be able to quantify the magnitude and distribution of reflected light. The magnitude and distribution of the reflected light depends on the total laser input power as well as its distribution, the laser incidence angle (angle between the normal to the transparent part surface and the laser beam), the laser light polarization as well as the surface and optical properties of the transparent part. A novel technique based on thermal imaging of the reflected light was previously developed by the authors. It is used in this study to characterize the magnitude and distribution of reflected light from thermoplastics as a function of thickness (1–3.1mm), laser incidence angle (20–40°) and surface roughness (0.04–1.04μm). Results from reflection tests on nearly polished nylon 6 (surface roughness between 0.04 and 0.05μm) have shown that, for the various thicknesses tested (1–3.1mm), the total reflection was larger than the specular top surface reflection predicted via the Fresnel relation. From these observations, it is conjectured that, in addition to top surface reflection, the bulk and/or bottom surface also contribute to the total reflection. The results also showed that reflection decreased slightly with increasing thickness. As expected, for the p-polarized light used in this study, the reflection decreased with increasing angle of incidence for the range of angles studied. It was also found that when the surface roughness was close to zero and when it was close to the wavelength of the input laser beam (i.e. 940nm), the reflectance values were close and reached a minimum between these two roughness values.