This study aims to explore the differences in sustained attention load caused by functional partitioning and object distribution on desktops, as measured by functional near-infrared spectroscopy (fNIRS), providing a basis for optimizing study/office desktop layouts. Three desktop configurations were used: L-shaped, linear, and U-shaped layouts. A total of 33 children (16 boys and 17 girls) participated in this study. All participants were instructed to maintain their regular sleep schedule and obtain ≥9 h of sleep per night prior to the experiment, and they completed the task under three desktop scenarios. As task performance during Go and No-Go trials was monitored, changes in prefrontal oxygenated hemoglobin (oxyHb) concentration in the prefrontal cortex were simultaneously measured, and hemodynamic responses were quantified using fNIRS-derived oxyHb changes as an index of cortical activation, thereby enabling investigation of the hemodynamic mechanisms underlying task performance. Changes in oxyHb concentration in the prefrontal cortex(PFC) were simultaneously measured. Repeated measures ANOVA was then performed on the mean oxyHb and cognitive performance indices. Results showed that the L-shaped desktop significantly reduced sustained attention load by providing a semi-enclosed, low-interference visual field; the linear desktop was second best due to its vertical separation; and the U-shaped desktop, while offering the largest space, contained the most redundant visual information, making it the least conducive to maintaining prolonged focus. These findings provide new fNIRS evidence for elucidating the role of the desktop microenvironment in sustained attention.