Aims. We present cosmological constraints from a joint analysis of type la supernova (SN Ia) observations obtained by the SDSS-II and SNLS collaborations. The dataset includes several low-redshift samples (z < 0.1), all three seasons from the SDSS-11 (0.05 < z < 0.4), and three years from SNLS (0.2 < z < 1), and it totals 740 spectroscopically confirmed type la supernovae with high quality light curves. Methods. We followed the methods and assumptions of the SNLS three-year data analysis except for the following important improvements: I) the addition of the full SDSS-II spectroscopically-confirmed SN la sample in both the training of the SALT2 light-curve model and in the Hubble diagram analysis (374 SNe); 2) intercalibration of the SNLS and SDSS surveys and reduced systematic uncertainties in the photometric calibration, performed blindly with respect to the cosmology analysis; and 3) a thorough investigation of systematic errors associated with the SALT2 modeling of SN la light curves. Results. We produce recalibrated SN la light curves and associated distances for the SDSS-II and SNLS samples. The large SOSS-II sample provides an effective, independent, low -z anchor for the Hubble diagram and reduces the systematic error from calibration systematics in the low -z SN sample. For a flat ACDM cosmology, we find Omega(m), = 0.295 0.034 (stat+sys), a value consistent with the most recent cosmic microwave background (CMB) measurement from the Planck and WMAP experiments. Our result is 1.8 sigma (stat+sys) different than the previously published result of SNLS three-year data. The change is due primarily to improvements in the SNLS photometric calibration. When combined with CMB constraints, we measure a constant dark energy equation of state parameter omega = -1.018 +/- 0,057 (sral+sys) for a fiat universe. Adding baryon acoustic oscillation distance measurements gives similar constraints: omega = 59 -1.027 0.055. Our supernova measurements provide the most stringent constraints to date on the nature of dark energy.