Monitoring human respiratory pattern has gained much attention among researchers across the current decade. Herein, we have used a novel graphene-based flow sensor as an airflow sensor to capture the behaviour of human respiration. The sensor takes advantage of a maze-like structure of vertically grown graphene nanosheets (VGNs) integrated with polydimethylsiloxane (PDMS) elastomer (25:1 weight ratio of prepolymer and curing agent). The VGNs/PDMS nanocomposite was characterised by exposing to a range of airflow rates (20-130 L/min). The sensor demonstrates a linear performance with high sensitivity and low response time (mostly below 1 s). To evaluate the experimental results and better understand flow dynamics around the sensor, a finite element simulation was conducted in COMSOL Multiphysics package. The piezoresistive effect of VGNs/PDMS thin film and fluid-solid interaction was thoroughly studied, and numerically computed sensor output in response to various flow rates was verified by experimental results. As a potential application, the VGNs/PDMS airflow sensor can detect the respiration pattern of human exercises like walking, jogging and running. The sensor shows a stable pattern with very low response and recovery time (1 s) during five consecutive cycles.