Surface hardness is a critical parameter to characterize steel properties. Its monitoring plays a key role in Industry. In the present paper, the spectral line intensities between element Fe Ⅰ 404.58 nm and Mn Ⅰ 403.07 was compared with 6 different hardness samples by using laser induced breakdown spectroscopy (LIBS) in self-built device in which laser energy was 30 mJ. The coefficient of determination for Mn (0.964) was better than Fe (0.916). The ratio of the Ferrum ionic to atomic spectral lines intensities(Fe Ⅱ/Fe Ⅰ)and laser induced plasma temperature have been characterized by the hardness of D2 steel for different heat treatments. The relationship has been obtained between surface hardness and the ratio of the Ferrum ionic to atomic spectral lines intensities (the coefficient of determination was 0.964). Fe Ⅱ 275.57 nm and Fe Ⅰ 276.75 nm were selected as analytical spectral lines. The relationship between surface hardness and laser induced plasma temperature has been established too (the coefficient of determination was 0.977). The hardness of D2 steel could be changed by different heat treatments, as well as the addition of alloying elements. For example, adding manganese into D2 steel could improve the hardness by refining grains. The correlation between Rockwell hardness of D2 steel for different content of manganese and Spectral signal intensity (Mn Ⅰ 403.07 nm) was established. The hardness of D2 steel does not increase monotonically with the increase of Mn content, but the spectral intensity changes consistently with the hardness. The experimental results validated that the element Fe and Mn, the Fe Ⅱ/Fe Ⅰ spectral lines intensities and laser induced plasma temperature had a good linear correlation with the hardness of D2 steel. And the different hardness obtained by different addition of Mn in D2 steel had a good linear correlation with a spectral intensity which verified the relation between hardness and spectral intensity. The determination of the surface harness via LIBS shows the feasibility of using LIBS as a reliable method for in situ industrial application for production control.