Esophageal histology is critical for diagnosis and surveillance of disease activity in eosinophilic esophagitis (EoE). A validated noninvasive biomarker has not been identified. We aimed to determine the utility of blood and urine eosinophil-associated proteins to diagnose EoE and predict esophageal eosinophilia. Blood and urine were collected from children undergoing endoscopy with biopsy. Absolute eosinophil count (AEC), plasma eosinophil-derived neurotoxin (EDN), eosinophil cationic protein (ECP), major basic protein-1 (MBP-1), galectin-10 (CLC/GAL-10), Eotaxin-2 and Eotaxin-3, and urine osteopontin (OPN) and matrix metalloproteinase-9 (MMP-9) were determined. Differences were assessed between EoE and control, and with treatment response. The capacity to predict EoE diagnosis and esophageal eosinophil counts was assessed. Of 183 specimens were collected from 56 EoE patients and 15 non-EoE controls with symptoms of esophageal dysfunction; 33 EoE patients had paired pre- and post-treatment specimens. Plasma (CLC/GAL-10, ECP, EDN, Eotaxin-3, MBP-1) and urine (OPN) biomarkers were increased in EoE compared to control. A panel comprising CLC/GAL-10, Eotaxin-3, ECP, EDN, MBP-1, and AEC was superior to AEC alone in distinguishing EoE from control. AEC, CLC/GAL-10, ECP, and MBP-1 were significantly decreased in patients with esophageal eosinophil counts <15/hpf in response to treatment. AEC, CLC/GAL-10, ECP, EDN, OPN, and MBP-1 each predicted esophageal eosinophil counts utilizing mixed models controlled for age, gender, treatment, and atopy; AEC combined with MBP-1 best predicted the counts. We identified novel panels of eosinophil-associated proteins that along with AEC are superior to AEC alone in distinguishing EoE from controls and predicting esophageal eosinophil counts.