Landscape Architects and allied professions are steadily integrating artificial intelligence (AI) and machine learning (ML) in practice and deployment to enhance design processes, optimize project management, and augment analytical practices. The application of Retrieval-Augmented Generation (RAG) models in the fields of landscape architecture, planning, ecology, and architecture is still an emerging area and is not yet fully understood or widely explored. RAG models integrate a pretrained language model with a retrieval system, effectively merging the processes of information retrieval and language generation into a cohesive framework. However, validation of information from large language models (LLMs) in Question-Answering Systems (QAS) (driven by AI/ML algorithms), such as ChatGPT and Google Gemini poses a challenge for landscape architects. The objective of this study was to assess the performance of the RAG model applied to landscape architecture literature. To address this objective, we developed a closed-domain neural network using open-source models trained on one issue of The Journal of Digital Landscape Architecture. To evaluate its performance, we queried the neural network on a series of landscape architectural tasks including design, theory, and analytical tasks. We then used quantitative measures to evaluate the performance. The results of ROUGE scores for the RAG demonstrate its effectiveness in capturing key concepts within the landscape architecture domain, particularly noting high precision values in Rouge-1 and Rouge-L metrics. While the model shows a lower performance in capturing two-word combinations as indicated by Rouge-2 scores, it successfully retrieved relevant information efficiently, as demonstrated by higher precision across other metrics. The study highlights the potential of Closed Domain Question Answering (CDQA) systems integrated with a RAG model, trained on specialized datasets, to enhance landscape architects' workflows. It also underscores the necessity of addressing challenges such as data curation, bias, and creativity limitations to maximize the utility and success of these tools in professional landscape architecture practice.