The significant rise in the complexity and scope of construction projects led by the introduction of highly advanced building systems characterizes the current construction industry. This entails a significant increase in coordination and planning and an overall change in the management culture of all project participants. On the other hand, project stakeholders are faced with an increasing demand from project owners to implement fast-track programmes to achieve an early return on investment. Consequently, claims and disputes throughout the majority of project delivery systems have surged, influenced by the project parties' inability to effectively manage the claims process. The aim of this paper is to introduce a new methodology for the automatic text classification of project delay claims documents to enhance efficiency in the management of delay claim process. The proposed model utilizes activity and Work Breakdown Structure keywords of a given delay event activity path for the training of the proposed model, which is then utilized to predict unlabeled project documents. The proposed model has been implemented on a series of delay claims events in a mega project, the implementation yielded promising results in the performance evaluation measures (precision, recall, and F1-Score) compared to similar text classification models