Abstract：The stability of suspensions is critical to the smooth operation of the entire production system, especially during material storage, transportation and production. Nevertheless, the challenge of achieving cost-effective and real-time online monitoring remains, owing to the variability in the particles size and concentration. In this study, a triboelectric probe integrated with deep learning technology were employed to achieve self-powered, real-time, online monitoring of the particles size and concentration, whose variations would influence the distribution of surface charges, thereby inducing changes in the macroscopic dielectric constant of the suspension. Such triboelectric probes successfully captures output signals that reflect changes in the properties of the suspension, based on the coupled effects of contact electrification and electrostatic induction at the liquid-solid interface. Furthermore, a convolutional neural network model within deep learning technologies were established for handling the relevant signals, and performed an average recognition accuracy exceeding 98% for both the particles size and concentration. The integration of the triboelectric probe and deep learning technology presents a novel approach for real-time online monitoring of particles, holds significant implications for the production security monitoring in industries such as pharmaceuticals, chemical engineering, and food production.