Lyme neuroborreliosis (LNB), a disseminated manifestation of Lyme borreliosis (LB), arises when Borrelia burgdorferi sensu lato (Bbsl) spirochetes disseminate within the host and damage the peripheral nervous system and meninges, and in rare cases, also the parenchyma of the central nervous system (CNS). While early-stage LB is diagnosed clinically, accurate diagnosis of LNB requires cerebrospinal fluid (CSF) analysis, demonstrating pleocytosis and intrathecal synthesis of Bbsl-specific antibodies. There are, however, limitations in current LNB diagnostics, such as the unspecific nature of pleocytosis and post-treatment persistence of intrathecal antibodies necessitating the search for novel biomarkers. In this study, we employed untargeted ultrahigh-performance liquid chromatography coupled with tandem mass spectrometry (UHPLC-MS/MS) to profile small metabolites (<1500 Da) in CSF samples from subjects with definite acute LNB. Comparative analyses of metabolite profiles were performed between pretreatment subject samples (n = 63) and the following groups: (A) a subset of samples collected three weeks after treatment initiation from the same individuals (n = 36), (B) Bbsl antibody-negative subjects (non-LNB, n = 61), (C) subjects with other CNS infections (n = 21). Additionally, pretreatment LNB samples were compared between individuals with radiculitis (n = 40) and those without radiculitis (n = 23) (D). Out of 4222 molecular features (MFs) detected, 131 were prioritized based on statistical significance and magnitude of change for further detailed structural characterization. Altered metabolite classes included compounds from lysophospholipids [e.g., lysophosphatidylcholine (16:0), and lysophosphatidylethanolamine (18:0)], sphingomyelins [e.g., sphingomyelin (d18:1/14:0) and sphingomyelin d16:1/16:0)], sphingoid bases (e. g., d19:0 sphinganine, and 3-ketosphingosine), fatty acid amides (e.g., palmitoleamide and oleamide), cyclic phosphatidic acids [i.e., cyclic phosphatidic acid (16:0) and cyclic phosphatidic acid (18:2)], and amino acid metabolism (i.e., DL-glutamine, 5-hydroxytryptophan and DL-kynurenine). These findings underscore the potential of CSF metabolomics as a powerful complementary tool for diagnosing LNB and differentiating it from other CNS conditions. The identified metabolic signatures offer a foundation for future biomarker development and may enhance diagnostic precision, guide treatment strategies, and deepen our understanding of LNB pathogenesis.