The aim of this study is to discuss the experimental performance analysis and deep learning based modelling of moist sodium polyacrylate material (also known as Orbeez) in a closed-loop heat pump dryer using R-134a as a secondary fluid. The experiments are performed on different weights of Orbeez at a constant air flow rate to calculate different performance parameters like coefficient of performance of the heat pump, drying rate, heat transfer rate by condenser, moisture extraction rate, and specific moisture extraction rate. The higher test loads such as 6, 7, and 8 kg are found better in terms of maximum coefficient of performance (5.2–5.8) and heat transfer rate (0.56–0.64 kW). Similarly, the higher test loads such as 6, 7, and 8 kg yield the highest moisture extraction rate (∼0.66–0.75 kg/h), specific moisture extraction rate (∼2.15–2.27 kg/kWh), and weight reduction (∼91%). The water removal rate depends on the moisture diffusivity, and it increases with an increase in the drying air temperature and drying air velocity. In addition, a deep learning model considering the most influential dryer inlet conditions (air temperature, air relative humidity, and airflow rate), time, and weight as the input features to estimate the dryer outlet conditions and weight reduction for assessment of drying kinetics of the considered material. A high accuracy (coefficient of determination = 0.997) makes it a simple, cost effective, and reliable method to predict the drying performance of various materials with a closed loop heat pump dryer.