We present a study of 33 Kepler planet-candidate host stars for which asteroseismic observations have sufficiently high signal-to-noise ratio to allow extraction of individual pulsation frequencies. We implement a new Bayesian scheme that is flexible in its input to process individual oscillation frequencies, combinations of them, and average asteroseismic parameters, and derive robust fundamental properties for these targets. Applying this scheme to grids of evolutionary models yields stellar properties with median statistical uncertainties of 1.2 per cent (radius), 1.7 per cent (density), 3.3 per cent (mass), 4.4 per cent (distance), and 14 per cent (age), making this the exoplanet host-star sample with the most precise and uniformly determined fundamental parameters to date. We assess the systematics from changes in the solar abundances and mixing-length parameter, showing that they are smaller than the statistical errors. We also determine the stellar properties with three other fitting algorithms and explore the systematics arising from using different evolution and pulsation codes, resulting in 1 per cent in density and radius, and 2 per cent and 7 per cent in mass and age, respectively. We confirm previous findings of the initial helium abundance being a source of systematics comparable to our statistical uncertainties, and discuss future prospects for constraining this parameter by combining asteroseismology and data from space missions. Finally, we compare our derived properties with those obtained using the global average asteroseismic observables along with effective temperature and metallicity, finding excellent level of agreement. Owing to selection effects, our results show that the majority of the high signal-to-noise ratio asteroseismic Kepler host stars are older than the Sun.