To develop a split glomerular filtration rate (sGFR) prediction model based on unilateral renal volume parameters using three-dimensional (3D) computed tomography (CT) volumetry. Clinical data (age, sex, height, weight, serum creatinine level [sCr], and sGFR measured by 99mTc-diethylene triamine pentaacetic acid nuclear renal scintigraphy with the double plasma sample method) of 67 healthy renal donors and 111 patients with hydronephrosis admitted from April 2016 to September 2021 were analysed. The split renal parenchymal volume (sRPV) and split renal calyces and pelvis volume (sRCPV) of 67 unilateral donor left kidneys and 111 hydronephrotic kidneys were measured. Statistical analysis of these parameters was performed to develop and validate the sGFR prediction model. sRPV (p<0.001), sRCPV (p=0.012), age (p=0.015), serum creatinine level (p=0.004), and weight (p=0.006) were significantly associated with the measured sGFR and were included in the sGFR prediction formula, which was constructed as: 68.710+0.093×sRPV-0.041×sRCPV-0.228×W-0.219×A-14.432×sCr (r2=0.416; where A is age, W is weight). The paired difference of internal validation between the measured sGFR (42.34±13.71 ml/min/1.73 m2) and the sGFR estimated by the prediction model (41.46±8.99 ml/min/1.73 m2) was 0.879±11.475 ml/min/1.73 m2 (p=0.492) with a 95% confidence interval of the mean difference of ±2.54 ml/min/1.73 m2. The proposed model based on sRPV and sRCPV parameters could be used for estimating split renal function of healthy renal donors and patients with hydronephrosis to some extent. Further studies are required to evaluate and rectify the model.