New model of quantum neural nework able to solve classification problems is presented. It is based on the extention of the model of quantum associative memory [1] and also utilizes Everett’s interpretation of quantum mechanics [2]–[4]. For presented model not neural weights but quantum entanglement is responsible for associations between input and output patterns. Distributed form of queries permits the system to generalize. Spurious memory trick is used to control the number of Grover’s iterations which is necessary to transform initial quantum state into the state which can give correct classification in most measurements. Numerical modelling of counting problem illustrates model’s behavior and its potential benefits.