With advancements in gravitational wave detection technology, an increasing number of binary neutron star (BNS) merger events are expected to be detected. Due to the narrow opening angle of jet cores, many BNS merger events occur off-axis, resulting in numerous gamma-ray bursts (GRBs) going undetected. Models suggest that kilonovae, which can be observed off-axis, offer more opportunities to be detected in the optical/near-infrared band as electromagnetic counterparts of BNS merger events. In this study, we calculate kilonova emission using a three-dimensional semi-analytical code and model the GRB afterglow emission with the open-source Python package afterglowpy at various inclination angles. Our results show that it is possible to identify the kilonova signal from the observed color evolution of BNS merger events. We also deduce the optimal observing window for SVOM/VT and JWST/NIRCam, which depends on the viewing angle, jet opening angle, and circumburst density. These parameters can be cross-checked with the multi-band afterglow fitting. We suggest that kilonovae are more likely to be identified at larger inclination angles, which can also help determine whether the observed signals without accompanying GRBs originate from BNS mergers.