This paper proposes an innovative deep learning model based on a multimodal framework for efficient keyword spotting in Mongolian lead-type newspaper images. The model integrates visual and textual data to enhance keyword localization accuracy and robustness, supporting both Query-by-Example (QbE) and Query-by-String (QbS) tasks. To address homoglyphic heterogeneity in Mongolian graphemes (i.e., different characters sharing identical glyphs), an intermediate code mapping mechanism based on Unicode standardization is applied, which unifies morphologically identical characters into a normalized intermediate representation layer, achieving a QbS accuracy of 95.57%. Additionally, to tackle the contextual dependency of character forms in Mongolian cursive writing, learnable positional encoding is incorporated into the word vector embedding layer. This enables the model to dynamically capture morphological changes of characters at different word positions, further improving QbS accuracy to 95.69%. Experiments demonstrate the model’s robustness under noisy conditions, confirming its effectiveness for Mongolian lead-type newspaper retrieval. This study provides a highly practical and reliable deep learning solution for the field of Mongolian image retrieval, laying a solid foundation for further enhancing the digital retrieval capabilities of Mongolian books, newspapers, and other literary resources.