Motivated by the observed excess of the di-photon signal in Higgs searches, sigma(gamma gamma)/sigma(SM) similar or equal to 1.5, we argue that models with enhanced Gamma(h -> gamma gamma) alone are the most favorable scenarios when the latest LHC/Tevatron results are all taken into account. We study the phenomenology of a supersymmetric scenario of light stau that predicts a 125 GeV SM-like Higgs boson with enhanced di-photon decay through light stau loops. Since it is extremely challenging to search the Drell-Yan stau pair at the LHC due to the small production rate, we focus on the parameter space with enhanced production of inclusive stau pairs, in particular, via b (b) over bar fusion or gaugino pairs. We study its phenomenology in both pure leptonic tau tau(+/-)(l) channels and hadronic tau tagged tau(h) channels. We find the same-sign dilepton from (chi) over tilde (+/-)(1) (chi) over tilde (0)(2) -> tau(+/-)(l)tau(+/-)(l) + X may significantly improve the discovery potential with even 7-8 TeV LHC of 0(30 fb(-1)) data. In the case of hadronic tau pair, we use the final state j + tau(h)tau(h) + is not an element of(T) to search and find that even with the most optimistic region of M-2 similar to 200-300 GeV, it requires at least 50 fb(-1) data of 14 TeV LHC to reach a significance of 3.5 sigma. Therefore, we conclude it is difficult to claim discovery only through hadronic tau based on the data by the 2012 shut-down. 5 sigma reach for our most optimistic region then requires 100 fb(-1) data with 14 TeV running. (C) 2012 Elsevier B.V. All rights reserved.