Black hole–neutron star mergers are among the most promising gravitational–wave sources for ground–based detectors , and gravitational waves from black hole–neutron star mergers are expected to be detected in the next few years . Simultaneous detection of electromagnetic counterparts with gravitational wave provides rich information about the merger events . Among the possible electromagnetic counterparts from the black hole–neutron star merger , the emission powered by the decay of radioactive r–process nuclei , so called kilonova/macronova , is one of the best targets for follow–up observation . We derive fitting formulas for the mass and the velocity of ejecta from a generic black hole–neutron star merger based on recently performed numerical relativity simulations . We combined these fitting formulas with a new semi–analytic model for a black hole–neutron star kilonova/macronova lightcurve , which reproduces the results of radiation–transfer simulations . Specifically , the semi–analytic model reproduces the results of each band magnitude obtained by the previous radiation transfer simulations \citep 2014ApJ…780…31T within \sim 1 ~ { } { mag } . By using this semi–analytic model , we found that , at 400 ~ { } { Mpc } , the kilonova/macronova is as bright as 22 – 24 ~ { } { mag } for the cases with a small chirp mass and a high black hole spin , and > 28 ~ { } { mag } for a large chirp mass and a low black hole spin . We also apply our model to GRB130603B as an illustration , and show that a black hole–neutron star merger with a rapidly spinning black hole and a large neutron star radius is favored .