Objective Metabolomics, the study of global alterations in small metabolites, is a useful tool to look for novel biomarkers. Recently, we reported a reprogramming of the serum metabolomic profile by nuclear magnetic resonance (NMR) spectroscopy following treatment in lupus nephritis (LN). This study aimed to compare the urine excretory levels of citrate and acetate in patients with biopsy-proven LN before and six months after cyclophosphamide induction therapy and to evaluate their correlation with the Systemic Lupus Erythematosus Disease Activity Index 2K (SLEDAI 2K) and renal SLEDAI. Methods Urine obtained from LN patients ( N = 18, 16 female) at diagnosis and six months following induction therapy with cyclophosphamide and healthy controls (HC; N = 18, median age = 35 years, all female) were stored at –80°C. Metabolomic profiling was done using high resolution 800 MHz 1D ¹H NMR spectroscopy. The urinary ratio of metabolites was calculated as (metabolite×1000)/creatinine. Disease activity was measured using the SLEDAI. Metabolomic profiles were compared between groups and correlated with clinical parameters. Results Compared to HC, LN patients had significantly lower median urinary citrate/creatinine levels (LN = 18.26, range 12.80–27.62; HC = 107.7, range 65.39–138.4; p < 0.0001) which significantly increased after six months of cyclophosphamide treatment (51.05, range 11.51–170.2; p = 0.03). LN patients also differed from HC by having a higher mean urinary acetate/creatinine ratio (LN = 17.44, range 11.6–32.7; HC = 9.61, range 7.97–13.71; p = 0.054) with a non-significant fall in values after six months of treatment. The Area under curve for differentiating LN from HC for urinary citrate was 0.9136, and urinary acetate was 0.6883. The urinary acetate levels correlated with SLEDAI ( r = 0.337, p = 0.048). Urinary citrate levels correlated positively with C3 ( r = 0.362, p = 0.03) and negatively with urine protein/creatinine ( r = −0.346, p = 0.039). Conclusions Urinary citrate, which reflects dampened aerobic glycolysis and oxidative phosphorylation, improved significantly and is a potential non-invasive biomarker for diagnosis and monitoring treatment response in LN.