Galaxy clusters are the most massive, gravitationally bound systems with masses (~$10^{14} h^{-1} M_{\odot}$), and therefore the statistical properties can be a powerful probe of cosmology. However, in order for us to use the cluster statistics for cosmology, it is of critical importance to have a well-calibrated scaling relation of cluster observables (e.g. optical richness) with cluster mass including the contribution of dark matter. For this, cross-correlations of clusters with shapes of background galaxies, so-called stacked weak lensing, offers a promising way of calibrating the scaling relation, because the stacked weak lensing measures the average mass distribution around the clusters. In this talk we use the redMaPPer cluster samples, the currently largest, homogenous sample of massive clusters, constructed from the SDSS data in order to calibrate the scaling relation with the weak lensing measurements. In doing this, we use a cosmic “emulator” of the halo-mass cross-correlations that give predictions of the cluster mass profile and are constructed from a set of high-resolution N-body simulations, as a function of halo mass and redshift, where halos are the places hosting massive clusters. We show how the cosmic emulator is powerful to calibrate the cluster observable-halo mass scaling relation, and then argue how the method can be applied to the ongoing Subaru Hyper Surpime-Cam survey in order to attain the full potential for constraining cosmological parameters.