<p><i>Background </i>robotic-assisted (RA) pedicle screw placement has been increasingly adopted in lumbar spine surgery because of its potential to improve accuracy compared with conventional fluoroscopy-assisted (FA) techniques. However, vertebral level-specific differences remain insufficiently characterized. This study compared RA and FA lumbar pedicle screw placement focusing on level-specific and side-specific radiographic accuracy. <i>Methods</i> a retrospective single-center comparative cohort study was conducted using prospectively collected data. A total of 2065 lumbar pedicle screws were analyzed, including 1053 placed using RA and 1012 using conventional FA instrumentation. Screw accuracy was assessed postoperatively using the Gertzbein-Robbins classification, with Grade A considered optimal. Level-specific and side-specific analyses were performed for L1-L5. Fisher’s exact test, clustered logistic regression, and propensity score weighted sensitivity analyses were used for statistical comparisons. <i>Results</i> Grade A placement was achieved in 99.1% of RA screws (1044/1053) versus 87.9% of FA screws (890/1012) (p &lt; 0.001). Robotic-assisted instrumentation demonstrated significantly higher rates of optimal screw placement across all lumbar levels and both sides. The greatest relative differences were observed at upper lumbar levels, particularly L1-L2. At L5, Grade A placement reached 100.0% on the right side and 99.5% on the left side in the RA cohort, compared with 89.2% and 88.1% in the FA cohort (both p &lt; 0.001). Fluoroscopy-assisted instrumentation independently predicted lower odds of optimal screw placement (OR 0.042, 95% CI 0.016–0.111; p &lt; 0.001). <i>Conclusions</i> robotic-assisted lumbar pedicle screw placement was associated with significantly higher radiographic Grade A screw placement accuracy than FA instrumentation.</p>