With the continuous development of informatization, the system operation in today's society is becoming increasingly efficient, which prompts individuals and teams to continuously optimize their workflow design and improve the efficiency of workflow modeling processes. To address these issues, we deeply explore the attribute information of workflows and apply it to workflow recommendations. This article proposes an enhanced Structure Aware Graph Embedding method (ESAGE) applied to workflow recommendation, which utilizes the node attribute information of the workflow graph to enhance the graph structure information, generate feature embeddings for the workflow to calculate the similarity between workflows. Based on this, we have achieved efficient recommendation in the workflow. In addition, this method is also applicable to generating feature embeddings for nodes in other graph-structured data. Finally, based on real datasets, we designed and conducted experiments, and the experimental results showed the effectiveness of this method in the practical process.