The aim of this work is to provide clues for the understanding of the star formation history of the universe. We study the process of star formation in galactic objects in hierarchical clustering scenarios using hydrodynamical cosmological simulations of standard Cold Dark Matter Universe. We estimate the specific star formation rate (sSFR) of galactic haloes, which is defined as the SFR per unit luminosity; their luminosities are obtained from the evolutionary synthesis population models of Charlot and Bruzual (1991). To confront the results of the simulations with observations, we consider the sample of field galaxies at high redshift up to z = 1.3 of the Canada-France Redshift Survey (CFRS, Hammer et al.,1997). The sSFR for these galaxies were estimated from their spectral properties. From the confrontation of models and observations, we see that galactic objects formed in hierarchical clustering scenarios seem to reproduce the specific SFR observed at intermediate (0.35 < z < 0.7) and high (0.7 < z < 1.3) redshifts in the CFRS and also their dependence with mass and redshift. Our results suggest that in both simulations and observations, the mechanisms controlling the transformation of gas into stars is more effective at high redshift and in smaller objects producing larger sSFR in low lµ-luminosity galaxies. We find that the sSFR depends on the redshift (Cora et al.,2000, in prep.). Starbursts tidally induced during mergers might be responsible for this trend.