Endometriosis is a chronic systemic inflammatory disorder characterized by endometrial tissue outside the uterine cavity in females. So far, the invasive laparoscopic method is the only gold standard diagnostic option for endometriosis. The study aims to develop a non-invasive diagnosis of endometriosis in patients presenting with one or more symptoms by analyzing the peripheral circulating miRNAs from the serum of the patients. A panel of miRNA 125b-5p, 342-3p, and Let-7b was developed to diagnose endometriosis. We performed the demographic profiling in 56 patients eliciting one or more symptoms of endometriosis without imaging evidence and compared them with 40 patients with the laparoscopically established endometriotic condition. Patients presenting with one or more of the clinical symptoms of endometriosis (n=56) served as the study group, and patients who were proven to be endometriotic by laparoscopy (n=40) served as the training (control) group. The fasting peripheral blood sample was collected, serum was separated, and cryo-preserved. qRT-PCR analysis of the selected miRNAs (miR 125b-5p, miR 342-3p, and Let-7b) was studied in both training (control) and study groups. The results were analyzed using a random forest (RF) approach machine learning algorithm and 10-fold cross-validation. Results indicate significant upregulation of miRNA125b-5p and miRNA 342-3p and downregulation of Let- 7b (p <0.001). Further, the samples derived from the study group with one or more symptoms of endometriosis and the proven endometriotic patients exhibited similar dysregulation of selected miRNAs (miR 125b-5p, miR 342-3p, and Let-7b). Based on our study, we propose the miRNAs panel consisting of miR 125b-5p, miR 342-3p, and Let-7b to be used as an early non-invasive diagnostic marker for endometriosis efficiently without the patients being subjected to invasive laparoscopy, which will reduce the time taken for diagnosis and commence the treatment earlier and prevent morbidity.