We here propose a new model-independent technique to overcome the circularity problem affecting the use of Gamma-Ray Bursts ( GRBs ) as distance indicators through the use of E _ { p } – E _ { iso } correlation . We calibrate the E _ { p } – E _ { iso } correlation and find the GRB distance moduli that can be used to constrain dark energy models . We use observational Hubble data to approximate the cosmic evolution through Bézier parametric curve obtained through the linear combination of Bernstein basis polynomials . In so doing , we build up a new data set consisting of 193 GRB distance moduli . We combine this sample with the supernova JLA data set to test the standard \Lambda CDM model and its w CDM extension . We place observational constraints on the cosmological parameters through Markov Chain Monte Carlo numerical technique . Moreover , we compare the theoretical scenarios by performing the AIC and DIC statistics . For the \Lambda CDM model we find \Omega _ { m } = 0.397 ^ { +0.040 } _ { -0.039 } at the 2 \sigma level , while for the w CDM model we obtain \Omega _ { m } = 0.34 ^ { +0.13 } _ { -0.15 } and w = -0.86 ^ { +0.36 } _ { -0.38 } at the 2 \sigma level . Our analysis suggests that \Lambda CDM model is statistically favoured over the w CDM scenario . No evidence for extension of the \Lambda CDM model is found .