The Standard Model of particle physics successfully explains the majority of experimental high energy physics data. The masses of the W and Z, the vector bosons of the weak interaction, have their origin in a spontaneous breaking of the gauge symmetry. This symmetry breaking is performed, using the Brout-Englert-Higgs mechanism, by introducing a new scalar field, whose quantum, the Brout-Englert-Higgs boson, has still to be experimentally observed. Direct searches at LEP have excluded the BEH boson with masses lower than 114 GeV/c2,and direct searches at the Tevatron have led to an exclusion of masses between 158 and 173 GeV/c2. The fit of precision electroweak measurements constrains the BEH boson mass to be less than 158 GeV/c2 (all these limits are at the 95% confidence level). The H → γγ decay channel provides a clean topology : consequently, despite its small branching ratio, this channel is one of the most promising for a BEH boson lighter than 130 GeV/c2 at LHC. Because of the narrow BEH resonance, the discovery potential depends strongly on the photon energy resolution : I will describe the photon reconstruction in the CMS detector and I will detail the corrections applied to the photon energy in order to improve the resolution.Neutral pions, which can decay in two photons, can be easily misidentified as photons. I will explain how some variables describing the shape of the energy deposit in the CMS electromagnetic calorimeter can be exploited, with the help of an artificial neural network, to discriminate neutral pions from photons. Then, I will evaluate the impact of the improvement on photon identification, using a cut on the neural network output, on the H → γγ analysis. Finally I will present the latest public results about the BEH searches at LHC with the 2011 data.