The recent measurement of the muon anomalous magnetic moment a µ shows a 2.6σ deviation from the standard model value. We show that it puts strong constraints on the parameter space of the two-Higgs doublet model (2HDM) II. The dominant contribution of the Higgs bosons comes at the two-loop level, and in order to explain the data it favors a pseudoscalar A with a light mass range and a large tanβ. At 95% C.L., the upper limit for m A is 29 (55) (85) GeV for tanβ = 30 (45) (60), and tanβ is bounded below at 17. This is in sharp contrast to the conclusion one draws from considering one-loop Higgs contributions alone. Finally, we also discuss the role of the Higgs contributions in the minimal supersymmetric standard model. * E-mail: cheung@phys.cts.nthu.edu.tw † E-mail: chouch@phys.sinica.edu. which may suggest the presence of contributions from physics beyond the SM. Taking the above numbers at face value, the range of ∆a µ at 95% C.L. (±1.96σ) is given by 10.3 × 10 −10 < ∆a µ < 74.9 × 10 −10 .Most of the extensions of the SM start with an extended Higgs sector, the simplest of which is the two-Higgs-doublet model (2HDM). In particular, the more interesting model II [3] shares the same Higgs structure as the minimal supersymmetric standard model (MSSM). Hence, going beyond the SM to look for extra contributions to a µ , the 2HDM should be among the first to be examined seriously.In this Rapid Communication, we investigate the contributions from the 2HDM II. Stringent constraints are thus obtained on the parameter space of the model. The results are summarized as follows. In the 2HDM II, the dominant contribution actually comes at the two-loop level, and in order to explain the data ∆a µ is preferred to come from the Higgs pseudoscalar A, the mass of which is required to be less than 29 (55) (85) GeV for tanβ = 30 (45) (60) by the 95% C.L. range of Eq.(2). Moreover, tanβ has to be larger than 17. Recently, there has been some work on the same subject [4]; however, the paper only considered the one-loop contributions from the Higgs bosons, which become smaller than their corresponding two-loop contributions when the Higgs bosons are heavier than a few GeV and substantially smaller for heavier Higgs bosons. It is important to note that the two-loop contributions considered here dramatically change the story of the Higgs sector contributions to a µ and invalidate most of the results of the one-loop studies. Although we focus on the 2HDM II here, a similar conclusion holds for most models with possible large contributions from the Higgs sector. In models with flavor-changing Higgs couplings, in particular, there are potentially 1-loop contributions substantially larger than the flavorconserving ones [5]. However, the Barr-Zee type 2-loop contributions would still have an important role to play and should be taken into consideration. This fact has often been overlooked in the literature.In the MSSM, the dominant contribution comes from the chargino-sneutrino-loop diagrams and it has been shown [6] that to sat...