Elementary (5 [approximately] 11 years old) students with mathematics difficulties often struggle with word-problem solving. This systematic review and network meta-analysis aimed to examine the comparative effectiveness of various instructional strategies and their combinations for improving word-problem solving skills in this population. We systematically searched for studies via Academic Search Complete, Education Source, Education Resources Information Center, Medline, PsycINFO, ProQuest Database, EdArXiv, and PsyArXiv by June 2024. We used a network meta-analytic approach to synthesize results and assess the risk of bias in the included studies. Based on 52 eligible studies involving over 6900 participants, the combination of concrete-semiconcrete-abstract instruction, graphic organizers, metacognitive strategies, and schema-based instruction was found to be the most effective intervention. In contrast, metacognitive strategies alone and the sole combination of CSA instruction, graphic organizers, and metacognitive strategies were the least effective. No single strategy emerged as most effective; rather, any successful combination required the inclusion of both concrete-semiconcrete-abstract and schema-based instruction. These findings support the "active ingredient" and synergistic effects hypotheses of multi-component interventions. For students with mathematics difficulties, a combination of concrete-semiconcrete-abstract and schema-based instruction consistently formed the foundational "active ingredient" with metacognitive strategies and graphic organizers acting as supportive elements that maximized the effectiveness of this foundational combination.