[en] Context: .One of the main goals of the COROT mission is to get precise photometric observations of selected bright stars in order to allow the modelling of their interior through asteroseismology. However, in order to interpret the asteroseismological data, the effective temperature, surface gravity, and chemical composition of the stars must be known with sufficient accuracy.Aims.To carry out this task, we have developed a spectroscopic method called APASS (Atmospheric Parameters and Abundances from Synthetic Spectra) which allows precise analysis of stars with a moderate to high rotational velocity, which is the case for most primary COROT targets.Methods.Our method is based on synthetic spectra in which individual lines are replaced by analysis units (isolated lines or line blends, depending on the crowding of the spectral region and on the rotational broadening). It works differentially with respect to the Sun and allows the atmospheric parameters and chemical abundances to be determined by considering analysis units with different sensitivities to these various parameters.Results.Using high signal-to-noise spectra and the APASS method, we determined the atmospheric parameters and chemical abundances of 13 primary COROT targets. Our results agree well with those obtained by Bruntt using his software VWA and with those obtained with the software TEMPLOGG. However, in both cases, our error bars are significantly smaller than those of other methods. Our effective temperatures are also in excellent agreement with those obtained with the IR photometry method. For five stars with relatively low rotational velocity, we also performed an analysis with a classical equivalent-width method to test agreement with APASS results. We show that equivalent-width measurements by Gaussian or Voigt profile-fitting are sensitive to the rotational broadening, leading to systematic errors whenever the projected rotation velocity is non-negligible. The APASS method appears superior in all cases and should thus be preferred.