Abstract Objectives The impact of glucocorticoids on neurocognitive performance in patients with SLE is not fully addressed. We aimed to study the effect of daily and cumulative glucocorticoid dose on neurocognitive performance in SLE patients. Methods Consecutive SLE patients and gender- and age-matched healthy controls (HCs) underwent the computer-based Automated Neuropsychological Assessment Matric (ANAM), which evaluates eight neurocognitive domains including learning, recall, visual perception, mental rotation, short-term memory, attention, sustained attention and working memory. The total and individual-domain throughput scores (TPSs) and the presence of cognitive dysfunction (total TPS <1.5 s.d. below the mean TPS of HCs) were compared between SLE patients and HCs. Within the SLE group, univariate and independent associations between prednisolone dose (daily and cumulative) and individual-domain TPS were studied by univariate and multivariable linear regression, respectively. Results A total of 96 SLE patients and 96 HCs were studied. SLE patients scored significantly worse across all the neurocognitive domains and had a significantly lower mean total TPS (P < 0.001) and a higher prevalence of cognitive dysfunction compared with HCs (25.0 vs 7.3%, P = 0.001). In SLE patients, daily prednisolone dose was significantly and negatively correlated with mathematical-processing TPS, which probes working memory (P = 0.018). No significant correlation between cumulative prednisolone dose and any of the individual-domain TPSs was found. In multivariable regression, higher daily prednisolone dose and doses >9 mg daily remained independently associated with lower mathematical-processing TPSs (P = 0.031). Conclusion Daily prednisolone dose ≥9 mg, but not cumulative glucocorticoid dose, had an independent negative impact on mathematical processing in SLE patients.