The effect of thin cirrus clouds in retrieving the dust optical depth from MODIS observations is investigated by using a simplified aerosol retrieval algorithm based on the principles of the Deep Blue aerosol property retrieval method. Specifically, the errors of the retrieved dust optical depth due to thin cirrus contamination are quantified through the comparison of two retrievals by assuming dust-only atmospheres and the counterparts with overlapping mineral dust and thin cirrus clouds. To account for the effect of the polarization state of radiation field on radiance simulation, a vector radiative transfer model is used to generate the lookup tables. In the forward radiative transfer simulations involved in generating the lookup tables, the Rayleigh scattering by atmospheric gaseous molecules and the reflection of the surface assumed to be Lambertian are fully taken into account. Additionally, the spheroid model is utilized to account for the nonsphericity of dust particles in computing their optical properties. For simplicity, the single-scattering albedo, scattering phase matrix, and optical depth are specified a priori for thin cirrus clouds assumed to consist of droxtal ice crystals. The present results indicate that the errors in the retrieved dust optical depths due to the contamination of thin cirrus clouds depend on the scattering angle, underlying surface reflectance, and dust optical depth. Under heavy dusty conditions, the absolute errors are comparable to the predescribed optical depths of thin cirrus clouds.