The global issue of pollution caused by endocrine-disrupting chemicals (EDCs) has been gaining increasing attention. Among the EDCs of environmental concern, 17β-estradiol (E2) can produce the strongest estrogenic effects when it enters the organism exogenously through various routes and has the potential to cause harm, including malfunctions of the endocrine system and development of growth and reproductive disorders in humans and animals. Additionally, in humans, supraphysiological levels of E2 have been associated with a range of E2-dependent disorders and cancers. To ensure environmental safety and prevent potential risks of E2 to human and animal health, it is crucial to develop rapid, sensitive, low cost and simple approaches for detecting E2 contamination in the environment. A planar microwave sensor for E2 sensing is presented based on the integration of a microstrip transmission line (TL) loaded with a Peano fractal geometry with a narrow slot complementary split-ring resonator (PF-NSCSRR) and a microfluidic channel. The proposed technique offers a wide linear range for detecting E2, ranging from 0.001 to 10 mM, and can achieve high sensitivity with small sample volumes and simple operation methods. The proposed microwave sensor was validated through simulations and empirical measurements within a frequency range of 0.5-3.5 GHz. The E2 solution was delivered to the sensitive area of the sensor device via a microfluidic polydimethylsiloxane (PDMS) channel with an area of 2.7 mm2 and sample value of 1.37 µL and measured by a proposed sensor. The injection of E2 into the channel resulted in changes in the transmission coefficient (S21) and resonance frequency (Fr), which can be used as an indicator of E2 levels in solution. The maximum quality factor of 114.89 and the maximum sensitivity based on S21 and Fr at a concentration of 0.01 mM were 1746.98 dB/mM and 40 GHz/mM, respectively. Upon comparing the proposed sensor with the original Peano fractal geometry with complementary split-ring (PF-CSRR) sensors without a narrow slot, several parameters were evaluated, including sensitivity, quality factor, operating frequency, active area, and sample volume. The results showed that the proposed sensor exhibited an increased sensitivity of 6.08% and had a 40.72% higher quality factor, while the operating frequency, active area, and sample volume showed decreases of 1.71%, 25%, and 28.27%, respectively. The materials under tests (MUTs) were analyzed and categorized into groups using principal component analysis (PCA) with a K-mean clustering algorithm. The proposed E2 sensor has a compact size and simple structure that can be easily fabricated with low-cost materials. With the small sample volume requirement, fast measurement with a wide dynamic range, and a simple protocol, this proposed sensor can also be applied to measure high E2 levels in environmental, human, and animal samples.