Developing models, using mathematics, and constructing explanations are three practices essential for science inquiry learning according to education reform efforts, such as the Next Generation Science Standards (NGSS Lead States, 2013). However, students struggle with these intersecting practices, especially when developing and interpreting mathematical models (i.e., graphs and corresponding equations) of scientific phenomena and writing explanations that connect the mathematical representations to their scientific meaning. As such, students require targeted support to address their difficulties while they are working in science inquiry contexts. Thus, recent work with the Inquiry Intelligent Tutoring System (Inq-ITS) virtual lab environment focuses on the design and development of performance-based formative assessments with the goal of automatically assessing and supporting students on fine-grained components underlying mathematical modeling as well as exploring novel techniques for assessing explanation writing. ii The studies presented in this dissertation were conducted to evaluate the efficacy of the automated scaffolds in the Inq-ITS environment and examine how large language models could be leveraged to assess students' written explanations about their mathematical models in a way that provides formative data to teachers to guide their instruction. These findings inform how innovative technologies, such as virtual inquiry environments, can be used to assess and support NGSS-aligned mathematical modeling and explanation writing competencies in science inquiry contexts, thus preparing students for future success in mathematics and science.