In this work, a localized plasmon-based sensor is developed for para-cresol (p-cresol) - a water pollutant detection. A nonadiabatic <inline-formula> <tex-math notation="LaTeX">$40~\mu \text{m}$ </tex-math></inline-formula> of tapered optical fiber (TOF) has been experimentally fabricated and computationally analyzed using beam propagation method. For optimization of sensor’s performance, two probes are proposed, where probe 1 is immobilized with gold nanoparticles (AuNPs) and probe 2 is immobilized with the AuNPs along with zinc oxide nanoparticles (ZnO-NPs). The synthesized metal nanomaterials were characterized by ultraviolet-visible spectrophotometer (UV–vis spectrophotometer) and transmission electron microscope (HR-TEM). The nanomaterials coating on the surface of the sensing probe were characterized by a scanning electron microscope (SEM). Thereafter, to increase the specificity of the sensor, the probes are functionalized with tyrosinase enzyme. Different solutions of p-cresol in the concentration range of <inline-formula> <tex-math notation="LaTeX">$0~\mu \text{M}$ </tex-math></inline-formula> – <inline-formula> <tex-math notation="LaTeX">$1000~\mu \text{M}$ </tex-math></inline-formula> are prepared in an artificial urine solution for sensing purposes. Different analytes such as uric acid, <inline-formula> <tex-math notation="LaTeX">$\beta $ </tex-math></inline-formula>-cyclodextrin, L-alanine, and glycine are prepared for selectivity measurement. The linearity range, sensitivity, and limit of detection (LOD) of probe 1 are <inline-formula> <tex-math notation="LaTeX">$0~\mu \text{M}$ </tex-math></inline-formula> – <inline-formula> <tex-math notation="LaTeX">$700~\mu \text{M}$ </tex-math></inline-formula>, 7.2 nm/mM (accuracy 0.977), and <inline-formula> <tex-math notation="LaTeX">$59.90~\mu \text{M}$ </tex-math></inline-formula>, respectively; and for probe 2 are <inline-formula> <tex-math notation="LaTeX">$0~\mu \text{M}$ </tex-math></inline-formula> – <inline-formula> <tex-math notation="LaTeX">$1000~\mu \text{M}$ </tex-math></inline-formula>, 5.6 nm/mM (accuracy 0.981), and <inline-formula> <tex-math notation="LaTeX">$57.43~\mu \text{M}$ </tex-math></inline-formula>, respectively. Thus, the overall performance of probe 2 is quite better due to the inclusion of ZnO-NPs that increase the biocompatibility of sensor probe. The proposed sensor structure has potential applications in the food industry and clinical medicine.