This survey provides a comprehensive review of large language models (LLMs) applied to graph learning, highlighting their potential in enhancing performance in graph-centric tasks such as link prediction and node classification. Despite the advancements in Graph Neural Networks (GNNs), challenges like data sparsity and limited generalization remain. LLMs, known for their strong language comprehension and summarization capabilities, have emerged as a promising solution to these issues. The survey introduces a novel taxonomy to categorize existing methods based on their framework design, focusing on four unique designs: GNNs as Prefix, LLMs as Prefix, LLMs-Graphs Integration, and LLMs-Only. Each category is detailed with representative examples, exploring their strengths and limitations. The survey also discusses future research avenues, including overcoming integration challenges and exploring new application areas. The goal is to serve as a valuable resource for researchers and practitioners interested in leveraging LLMs in graph learning, fostering continued progress in this dynamic field.This survey provides a comprehensive review of large language models (LLMs) applied to graph learning, highlighting their potential in enhancing performance in graph-centric tasks such as link prediction and node classification. Despite the advancements in Graph Neural Networks (GNNs), challenges like data sparsity and limited generalization remain. LLMs, known for their strong language comprehension and summarization capabilities, have emerged as a promising solution to these issues. The survey introduces a novel taxonomy to categorize existing methods based on their framework design, focusing on four unique designs: GNNs as Prefix, LLMs as Prefix, LLMs-Graphs Integration, and LLMs-Only. Each category is detailed with representative examples, exploring their strengths and limitations. The survey also discusses future research avenues, including overcoming integration challenges and exploring new application areas. The goal is to serve as a valuable resource for researchers and practitioners interested in leveraging LLMs in graph learning, fostering continued progress in this dynamic field.