Recently, the CMS and ATLAS collaborations have announced the results for H → Z[→ ℓ+ℓ−]γ with ℓ = e or μ [1, 2], where H → Zγ is a sub-process of H → ℓ+ℓ−γ. This semi-leptonic Higgs decay gets both resonant and non-resonant contributions from the loop-induced H → Z[→ ℓ+ℓ−]γ. To probe further features coming from these contributions to H → ℓ+ℓ−γ, we argue that the polarization of the final state leptons is also an important parameter. We show that the interference of resonant and non-resonant contributions, which is negligible in the case of unpolarized leptons, plays a significant role when considering the polarization of the final state lepton. For this purpose, we have calculated the polarized decay rates and the longitudinal (PL), normal (PN), and transverse (PT) polarization asymmetries. We find that these asymmetries purely come from the loop contributions and are helpful to further investigate the resonant and non-resonant nature of H → Z[→ ℓ+ℓ−]γ decay. We observe that for ℓ = e, μ, the longitudinal decay rate is highly suppressed around mℓℓ ≈ 60 GeV when the final lepton spin is $$ -\frac{1}{2} $$
−
1
2
, dramatically increasing the corresponding lepton polarization asymmetries. Furthermore, we analyze another observable, the ratio of decay rates $$ {R}_{i\pm}^{{\mathcal{ll}}^{\prime }} $$
R
i
±
ll
′
, where ℓ and ℓ′ refer to different final state lepton generations. Precise measurements of these observables at the HL-LHC and the planned e+e− collider can provide fertile ground to test not only the SM but also to examine the signatures of possible NP beyond the SM.