Sequential recommendation aims to recommend items to users by taking into account the order of their previous interactions. Despite their success, most of the traditional approaches fail to capture a user’s dynamic context due to the limited availability of interaction sequences. Cross-domain recommendation (CDR) aims to enhance the recommendation quality in the target domain by harnessing data from several other diverse domains. However, a notable challenge in CDR is to accurately capture intra-sequence and inter-sequence item interactions across different domains. This paper tackles this challenge by simultaneously learning individual and cross-domain user preferences. We propose a lightweight Graph Neural Network (GNN) that exploits valuable intra-sequence and inter-sequence patterns. Our proposed lightweight GNN streamlines complexity while enhancing recommendation accuracy across diverse experimental scenarios. Extensive experiments corroborate our claim.