In this work we present a systematic method to fit and evaluate unknown optical parameters of a Monte-Carlo scintillator model based on different measurements. Monte-Carlo simulations are an important tool for the development of scintillation based radiation detector systems. It is usually used to understand how light propagates within the crystal which is especially important if one wants to analyze the impact of different photonic techniques to improve time resolution and light yield. The problem for most scintillator models is to gather a set of accurate optical parameters of the different components. While some parameters can be directly measured or taken from literature (like index of refraction, emission spectrum), some can only be measured with a complex and expensive setup and others are not be directly accessible or vary from crystal to crystal. A possibility to narrow down these parameters is the parameter fitting technique. It uses a set of measurements from a minimalistic setup to reduce other unknown parameters and uncertainties. Within the different measurements one or several parameters are changed, like the coupling medium or the crystal aspect ratio. The same is done in the MC model while varying the parameters one wants to fit. In contrast to related work our model additionally emphasizes on modeling the surface state of the crystals. For small crystals like used in medical imaging applications (usually <150mm3) the accurate modeling of the surface roughness is of great importance. In our simulations we could show, that for small crystals optical photons interacting between 20 and 30 times with the sidewalls before they get extracted. Therefore the surface state will affect the light transport to a large fraction and has to be taken into account. To fit our set of parameters, our technique uses a large set of different LY measurements obtained from different configurations of the same type of crystal. After the parameter fitting, the accuracy of our model is validated by comparing the simulation predictions to coincidence time resolution and angular distribution measurements.