In this paper, an organic macrocycle based on pillar[5]arene molecule including pyridine units (pyr-P5) was synthesized and its sensing ability was evaluated by integrating its spun thin films into Quartz Crystal Microbalance (QCM) and Surface Plasmon Resonance (SPR) sensing systems. The pyr-P5 mass sensitive-sensor's sensitivity and limit of detection (LOD) performance for benzene, toluene, and m-xylene were calculated to be 2.44, 1.67, and 1.11 Hz ppm−1 and 1.22, 1.79, and 2.70 ppm, respectively. In order to illuminate the swelling characteristics of the pyr-P5 optical sensor, the diffusion coefficients of these vapors were calculated by applying the early-time Fick's diffusion equation. The nonlinear autoregressive with exogenous input neural network was designed by utilizing experimental data from SPR kinetic results to model the change in photodetector response. The calculated diffusion coefficients using artificial neural network model are approximately equal to real-data diffusion coefficients as verified by very high correlation coefficients.