The Milky Way is only one out of billions of galaxies in the universe. However, it is a special galaxy because it allows to explore the main mechanisms involved in its evolution and formation history by unpicking the system star-by-star. Especially, the chemical fingerprints of its stars provide clues and evidence of past events in the Galaxy’s lifetime. These information help not only to decipher the current structure and building blocks of the Milky Way, but to learn more about the general formation process of galaxies. In the past decade a multitude of stellar spectroscopic Galactic surveys have scanned millions of stars far beyond the rim of the solar neighbourhood. The obtained spectroscopic information provide unprecedented insights to the chemo-dynamics of the Milky Way. In addition analytic models and numerical simulations of the Milky Way provide necessary descriptions and predictions suited for comparison with observations in order to decode the physical properties that underlie the complex system of the Galaxy. In the thesis various approaches are taken to connect modern theoretical modelling of galaxy formation and evolution with observations from Galactic stellar surveys. With its focus on the chemo-kinematics of the Galactic disk this work aims to determine new observational constraints on the formation of the Milky Way providing also proper comparisons with two different models. These are the population synthesis model TRILEGAL based on analytical distribution functions, which aims to simulate the number and distribution of stars in the Milky Way and its different components, and a hybrid model (MCM) that combines an N-body simulation of a Milky Way like galaxy in the cosmological framework with a semi-analytic chemical evolution model for the Milky Way. The major observational data sets in use come from two surveys, namely the “Radial Velocity Experiment” (RAVE) and the “Sloan Extension for Galactic Understanding and Exploration” (SEGUE). In the first approach the chemo-kinematic properties of the thin and thick disk of the Galaxy as traced by a selection of about 20000 SEGUE G-dwarf stars are directly compared to the predictions by the MCM model. As a necessary condition for this, SEGUE's selection function and its survey volume are evaluated in detail to correct the spectroscopic observations for their survey specific selection biases. Also, based on a Bayesian method spectro-photometric distances with uncertainties below 15% are computed for the selection of SEGUE G-dwarfs that are studied up to a distance of 3 kpc from the Sun. For the second approach two synthetic versions of the SEGUE survey are generated based on the above models. The obtained synthetic stellar catalogues are then used to create mock samples best resembling the compiled sample of observed SEGUE G-dwarfs. Generally, mock samples are not only ideal to compare predictions from various models. They also allow validation of the models' quality and improvement as with this work could be especially achieved for TRILEGAL. While TRILEGAL reproduces the statistical properties of the thin and thick disk as seen in the observations, the MCM model has shown to be more suitable in reproducing many chemo-kinematic correlations as revealed by the SEGUE stars. However, evidence has been found that the MCM model may be missing a stellar component with the properties of the thick disk that the observations clearly show. While the SEGUE stars do indicate a thin-thick dichotomy of the stellar Galactic disk in agreement with other spectroscopic stellar studies, no sign for a distinct metal-poor disk is seen in the MCM model. Usually stellar spectroscopic surveys are limited to a certain volume around the Sun covering different regions of the Galaxy’s disk. This often prevents to obtain a global view on the chemo-dynamics of the Galactic disk. Hence, a suitable combination of stellar samples from independent surveys is not only useful for the verification of results but it also helps to complete the picture of the Milky Way. Therefore, the thesis closes with a comparison of the SEGUE G-dwarfs and a sample of RAVE giants. The comparison reveals that the chemo-kinematic relations agree in disk regions where the samples of both surveys show a similar number of stars. For those parts of the survey volumes where one of the surveys lacks statistics they beautifully complement each other. This demonstrates that the comparison of theoretical models on the one side, and the combined observational data gathered by multiple surveys on the other side, are key ingredients to understand and disentangle the structure and formation history of the Milky Way.