We study the contribution of a generic charged Higgs (H + ) to the muon anomalous magnetic moment a µ with the SUSY soft breaking parameters. We find out that the deviation between the experimental data and the predicted SM value on a µ can be explained by the two-loop charged Higgs diagrams even with m H + ∼ 400 GeV .It is believed that the muon anomalous magnetic moment, a µ ≡ (g µ −2)/2, would provide precision tests of the standard model (SM) and probe for new physics [1]. Recently, it has been measured at BNL [2] with the data a exp µ = 116 592 023(151) × 10 −11 .(1)The experimental value in Eq. (1) differs from that in the SM [2] despite of the several different theoretical predictions from hadronic contributions [3,4]. In Ref.[2], it was reported that ∆a µ ≡ a exp µ − a SM µ = 426 ± 165 × 10 −11 ,while a recent calculation [4] based on a different estimation from the hadronic part gave ∆a µ = 375 ± 170 × 10 −11 .The values in Eqs. (2) and (3) indicate a window for new physics at 2.6σ and 2.2σ levels, which are translated into 215 × 10 −11 ≤ ∆a µ ≤ 637 × 10 −11 (90% CL) ,andrespectively. It is clear that both ranges in Eqs. (4) and (5) suggest the existence of new physics beyond the SM. However, one must caution about this less than 3σ result until the experiment of E821 at BNL is completed, which should increase the statistical significance at more than 6σ level [5], and the theoretical uncertainties from the hadronic part in a SM µ are further reduced. Recently, various models, such as those with SUSY, scalar bosons, and extra dimensions, which could lead to ∆a µ = O(400 × 10 −11 ) have been explored [1,6,7,8]. In particular, it is discussed extensively to use scalar Higgs bosons in SUSY-like theories as the viable candidates. In Refs. [7,8], the possibilities of using light neutral Higgs bosons with a large tanβ to account ∆a µ at the one-and two-loop levels were studied. It is known that a large tanβ is interesting theoretically since the unification of bottom and tau Yukawa couplings and the explanation of the top to the bottom mass ratio are realized in GUTs if tanβ ∼ O(50) [9]. With this possible large tanβ, it is found that the mass for the scalar boson has to be less than 5 GeV [7] and that for the pseudoscalar 75 GeV with including the Barr-Zee type [10] of the two-loop diagrams [8]. One may conclude that ∆a µ cannot arise from either a scalar or pseudoscalar in the minimal supersymmetric model (MSSM) due to the experimental limits on the scalar and pseudoscalar masses, which are in the ranges of 85 − 95 GeV [11,12,13].In this paper, we would like to examine whether it is possible to use a charged Higgs boson in SUSY models to induce ∆a µ beyond the one loop level. It is known that the challenge with a charged Higgs in theories is how to escape the constraint from the experimental value of B(B → X s γ) = 2.85 ± 0.41 × 10 −4 [14] which is consistent with that of 3.29 ± 0.33 × 10 −4 predicted in the SM. Following the analysis of [15,16] with the next-to-leading order (NLO) QCD corrections, ...