We investigated using a custom NGS panel of 149 genes the mutational landscape of 64 consecutive adult patients with tyrosine kinase fusion‐negative hypereosinophilia (HE)/hypereosinophilic syndrome (HES) harboring features suggestive of myeloid neoplasm. At least one mutation was reported in 50/64 (78%) patients (compared to 8/44 (18%) patients with idiopathic HE/HES/HEUS used as controls; p < .001). Thirty‐five patients (54%) had at least one mutation involving the JAK‐STAT pathway, including STAT5B (n = 18, among which the hotspot N642H, n = 13), JAK1 (indels in exon 13, n = 5; V658F/L, n = 2), and JAK2 (V617F, n = 6; indels in exon 13, n = 2). Other previously undescribed somatic mutations were also found in JAK2, JAK1, STAT5B, and STAT5A, including three patients who shared the same STAT5A V707fs mutation and features consistent with primary polycythemia. Nearly all JAK‐STAT mutations were preceded by (or associated with) myelodysplasia‐related gene mutations, especially in RNA‐splicing genes or chromatin modifiers. In multivariate analysis, neurologic involvement (hazard ratio [HR] 4.95 [1.87–13.13]; p = .001), anemia (HR 5.50 [2.24–13.49]; p < .001), and the presence of a high‐risk mutation (as per the molecular international prognosis scoring system: HR 6.87 [2.39–19.72]; p < .001) were independently associated with impaired overall survival. While corticosteroids were ineffective in all treated JAK‐STAT‐mutated patients, ruxolitinib showed positive hematological responses including in STAT5A‐mutated patients. These findings emphasize the usefulness of NGS for the workup of tyrosine kinase fusion‐negative HE/HES patients and support the use of JAK inhibitors in this setting. Updated classifications could consider patients with JAK‐STAT mutations and eosinophilia as a new “gene mutated‐entity” that could be differentiated from CEL, NOS, and idiopathic HES.