Rare loop-induced decays are sensitive to New Physics in many Standard Model extensions. In this paper we discuss the reconstruction of the radiative penguin decays b→sγ and the electroweak penguin decay b→ s at the LHC. The expected annual yields and B/S estimates are presented. Studies of the rare radiative penguin decays b→sγ, the electroweak penguin decay b→ s, and the decay B s →µµ [1] allow to extract valuable information about penguin and box loop-diagrams. The complex couplings of new particles may result in enhancement of decay rates or in the appearance of non-trivial CP-violating phases. For example for the decay B d →K * γ because of the one-diagram dominance (the strong phase appears only at order α S and 1/m b ) the direct CPasymmetry is reliably predicted in the SM to be ≤ 1% [2], but for some SUSY scenarios it could be as large as 10-40% [2,3].Due to the V − A structure of the weak current the photon polarisation in b → s (d) γ transitions is almost 100%. In the SM this causes mixinginduced CP-asymmetries to vanish [4], while in extensions of the SM these asymmetries could be as large as 50% [5]. This effect can be used as a probe for the spin structure of new particles.The test of QCD models in radiative penguin decays still plays an important rôle [6]. The ratiowith moderate theoretical uncertainty [7].The forward-backward asymmetry A FB for the decay b→ s, is defined through the angle θ FB between the + and the b hadron flight directions in the di-lepton rest frame. The shape of the asymmetry A FB m 2 and especially the position of the zero crossing in the SM are almost unaffected by hadronic form factor uncertainties, thus providing a good basis for searching for deviations [8].The ratio of b→µµs and b→ees decays in any exclusive mode is also a clean probe of the SM. Lepton-universality predicts this ratio to be 1 with theoretical errors below 1% [9]. The LHC will produce copious amounts of bhadrons, with a total bb cross-section of 500 µb. This potential will be exploited by the ATLAS, CMS and LHCb experiments.ATLAS and CMS are general-purpose central spectrometers designed for new physics searches at high luminosity [10]. Yet they will have a small trigger bandwidth dedicated to B-physics for decays involving muons during the initial running at lower luminosity. We assume for the following that this programme covers 3 years of running at L = 10 33 cm −2 s −1 , i.e. 30 fb −1 . LHCb is a forward spectrometer [11] optimised for b physics. Its main features are the precise vertex detector, the two RICH detectors and the versatile trigger with a 2 kHz output rate dominated by pp → bbX events. LHCb will operate at a lower luminosity of L = 2 · 10 32 cm −2 s −1 , corresponding to 2 fb −1 per year. The reconstruction of rare b decays at LHC is a challenge due to the small rates and large backgrounds from various sources. The most critical is the combinatorial background from pp → bbX events, containing secondary vertices and characterised by high charged and neutral multiplicities.