<p>Flexible neuromorphic electronics are attracting increasing attention as promising platforms for bio-integrated and wearable computing. However, achieving stable and reproducible synaptic functionality on soft substrates remains challenging. This difficulty mainly originates from material instability, fabrication complexity, and limited controllability of transient charge dynamics, which hinder faithful emulation of essential synaptic behaviors such as short-term plasticity (STP). Here, we present a hybrid synaptic transistor that integrates an inorganic HfOₓ charge-trap layer with a solution-processed DPP-DTT (PDPP2T-TT-OD) organic semiconductor channel. This hybrid architecture combines the stability and reliability of HfOₓ with the processability and mechanical compliance of DPP-DTT, enabling low-temperature fabrication and reliable integration on flexible substrates. The device reproduces key biological synaptic functions through a transient hole trapping/detrapping mechanism under gate bias, exhibiting paired-pulse facilitation, post-tetanic potentiation, and spike-rate-dependent plasticity. It also demonstrates reliable potentiation and depression under repetitive pulse stimulation. Leveraging its nonlinear dynamic response and fading-memory characteristics, the device is further implemented in a physical reservoir computing (PRC) framework, achieving high recognition accuracy in Modified National Institute of Standards and Technology (MNIST) digit classification while markedly reducing computational cost and training time. Additionally, its capability to process multi-channel biological time-series was successfully validated through a surface electromyography (sEMG)-based hand gesture recognition task. The proposed device can be a promising hybrid charge-trap synapse, serving as a fundamental building block for next-generation flexible neuromorphic system.</p>