Context retrieval and classification has always been an area of interest for researchers around the world. Data diversity is based on the type of information it contains, including text, image, video, and audio documents, as well as its sources, such as sensor data. In fact, this evolution of data is set to accelerate and even continue. To identify all the textual documents that match a query, an information retrieval (IR) system provides a method for verifying whether the words in each document correspond to the user’s query. The basic assumption of most systems is that the words extracted from textual documents are accurate and definitive. Sometimes, it is challenging to apply this assumption. This work proposes a new efficient model for indexing uncertain textual documents that are capable to reason in an environment with uncertain data. Specifically, the proposed method consists of two main phases: 1) to process uncertain data within a textual document, and 2) to determine a new approach of possibilistic syntactic indexing. These two phases are presented in the form of two algorithms. A series of experiments were carried out using different standard test data sets to evaluate the proposed work by comparing it with existing approaches.