Aiming at the accurate fluid control and the miniaturization of piezoelectric pump system, a flow self-measurement method for single chamber and single vibrator piezoelectric pump was proposed based on piezoelectric self-sensing and neural network. A bimorph vibrator was applied in piezoelectric pump acting as actuator and sensor simultaneously. Firstly, the influencing factors of flow were analyzed using the pi theorem of hydromechanics. Secondly, the relationship between sensing piezoelectric signal and deformation of the vibrator was studied. Then, the information of flow implicit in the sensing piezoelectric signal was found. Thus, a prototype of the parameter measurement circuit for sensing piezoelectric signal was made, and a BP neural network model for flow prediction was constructed by using the parameters acquired from the circuit as the input vector. Experimental results show that the correlation coefficient between the predicted data and the measured data is above 0.9993, and the maximum relative error is less than 3.46%. the results of prediction are close to the actual data. The method proposed for flow self measurement has a good accuracy.