The evolution of stellar disks is of great importance for understanding many aspects of galaxy formation. In this work we perform stellar population synthesis on radially resolved photometry of 564 disk galaxies from the SDSS DR5, selected to have both spectra of the central regions and photometry. To explore fully the multi-dimensional likelihood space defined by the output parameters of the spectral synthesis, we use Markov Chain Monte Carlo to quantify the expectation values, the uncertainties and the degeneracies of the parameters. We find good agreement between the parameter values obtained using the SDSS broad-band colors and the spectra respectively. In general the derived mean stellar age and the best-fit stellar metallicity decline in value from the galaxy center to the outer regions (around 1.5 half-light radii), based on sub-samples defined by concentration index. We also find that the radial dependency of the stellar population parameters exhibits a significant variation, and this diversity is likely related to morphology and the physics of star formation.