<p>Mathematics writing has received increasing attention in recent years; however, visuals—an important representation for constructing mathematical meaning—are often overlooked in this line of research. To address this gap, the present study introduces visual mathematics writing, a multimodal approach that integrates visual and textual representations to support students in expressing mathematical ideas more vividly and coherently. We conducted a comprehensive investigation of students’ visual mathematics writing and further examined how mathematical competence and general writing ability were associated with students’ performance across different visual mathematics writing features. Based on a sample of 140 fourth graders in China, our findings revealed that: (a) students performed better on visual than on textual features, and showed challenges in visual-text connection, (b) mathematical competence, rather than general writing ability, was most strongly associated with both textual and visual features; however, students’ performance on the visual–text connection feature remained consistently weak regardless of their levels of mathematical competence and general writing ability. These challenges appear to be more closely associated with students’ mathematical competence than with their general writing ability, and they likely require targeted instructional support to strengthen the integration between visuals and text.</p>