Graph Neural Networks (GNN) is a type of graph embedding approach designed to convert the nodes and edges in a graph into low-dimensional vector representations while preserving important structural and relational information. Node classification is the process in which the nodes of a graph are assigned labels according to the characteristics that are associated with those nodes. The objective of this paper is to perform node classification for Malayalam, an official language in India using various GNN approaches. Among all the GNN techniques used for node classification in Malayalam, the Relational Graph Convolutional Network (RGCNConv) model demonstrated the highest average accuracy of 70.87%. While this work may not be on par with the English research, our innovative outcome in Malayalam using GNN represents the first attempt in this area. Given the resource constraints associated with the Malayalam language, we have developed and made publicly available a node classification dataset that will serve as a foundation for future node classification research in Malayalam.