Recommendation systems create complex heterogeneous information networks (HINs) with a diversity of relational connections. Developing scalable recommendation systems to efficiently find similar users and items based on relations is a significant challenge. Hashing techniques can help mitigate scalability issue. Current relation-aware techniques focus mainly on a single metapath, which is a sequence of node and relation types. Using multiple metapaths, the system can capture various types of relationships between entities simultaneously. Therefore, it can create a richer and more accurate representation of user preferences, which improves the performance of recommendations. To address the issue that existing hashing approaches fail to effectively and efficiently exploit the potential of multiple relations, this paper introduces a novel Cross-Metapath Hashing model tailored for recommendation systems. This approach helps facilitate efficient neighborhood formation and recommendation making. The experiments conducted on a large-scale real-world dataset show that the proposed approach is effective.