This chapter explores the use of Explainable Artificial Intelligence (XAI) in the optimization of design parameters in surface plasmon resonance (SPR) sensors and the analysis of oil–water emulsion stability using tilted fiber Bragg grating (TFBG) sensors. These sensors are critical for applications such as biosensing, gas detection, environmental monitoring, and the petroleum industry etc. The study focuses on: (i) optimizing key design parameters for SPR sensors, including wavelength (λ) and metal layer thickness (dm) leading to maximum figure of merit (FOM), and (ii) analysing surfactant concentration (Cs) and rotational speed (RPM) for oil–water emulsion stability using TFBG sensors. Machine learning (ML) models and XAI methods like SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-Agnostic Explanations) are applied to SPR and TFBG sensors to enhance transparency and interpretability in the sensor design process. In the SPR sensor analysis, our findings demonstrate that dm has a significantly greater impact on optimizing the FOM than λ. Further, the findings are in corroboration with the fundamental concept of radiation damping in plasmonic structures. XAI techniques provided valuable insights into the contribution of these parameters, facilitating a more efficient and reliable sensor design. For oil–water emulsion stability, the study shows that Cs is the dominant factor influencing stability, with RPM and λ having comparatively smaller impacts. By utilizing ML and XAI techniques for SPR sensors and oil–water emulsion analysis, this chapter provides a comprehensive understanding of the key factors driving sensor performance and emulsion stability. The chapter concludes by suggesting future research directions, including the integration of advanced materials and real-time XAI applications to further improve the functionality and performance of these sensors. By implementing XAI methods such as SHAP and LIME alongside ML models like Gaussian Process Regression (GPR) and Random Forest, we gain deeper insights into the contributions of design parameters. The findings demonstrate that XAI enhances the reliability and interpretability of ML models, leading to more informed and trustworthy decision-making in photonic device design. This approach not only optimizes device performance but also aligns with the physical principles governing photonic systems, thereby bridging the gap between complex ML models and practical application.