We introduce a hybrid quantum classical neural network that can be trained to generate quantum states. The quantum neural network consists of quantum and classical layers. The fully connected quantum layer will be implemented by a quantum circuit for possible speedup over classical methods with linear quantum operation and the classical layer will be implemented by the usual classical non-linear activation functions. To demonstrate the power of the proposed new construction, we present results of using the quantum-classical hybrid neural network to generate eigenstates of simple molecules, complex entangled ground states and ground states of the transverse Ising Hamiltonian with varying local fields and spin-spin couplings. The approach achieved high accurate results and can be generalized for creating quantum states of complex systems.