Designing, optimizing and analyzing optical systems as part of the implementation process into production of modern luminaires require using advanced simulation and computational methods. The progressive miniaturization of LED (light emitting diode) chips and growth in maximum luminance values, achieving up to 10 8 cd/m 2 , require constructing very accurate geometries of reflector and lens systems producing complex luminous intensity distributions while reducing discomfort glare levels. Currently, the design process cannot function without advanced simulation methods. Today's simulation methods in the lighting technology offer very good results as far as relatively large conventional light sources such as halogen lamps, metal halide lamps and high pressure sodium lamps are concerned. Unfortunately, they often fail in the case of chip-on-board LED light sources whose luminous surface dimensions are increasingly often contained inside a cube of the side length below 1mm. With the high sensitivity of such small chips and lenses with dimensions ranging from a just a few to between 10 and 20 mm, which is presented in this paper, modern luminance distribution measurement methods, luminance modelling and ray tracing methods should be used to minimize any errors arising from incorrectly projecting the design in the final physical model. Also, very importantly, focus should be directed towards reducing a chance of making a mistake while collimating the position of the light source inside the optical system. The paper presents a novel simulation calculation method enriched with an analysis of optical system sensitivity to a light source position. The results of simulation calculations are compared with the results of laboratory measurements for corresponding systems.