Massive multiple-input and multiple-output (MIMO), especially in the millimeter-wave (mmWave) frequency band, has been recognized as a promising technology for future wireless communications. However, most previous research in mmWave massive MIMO only uses conventional MIMO channel model without taking into account the spatial-wideband effect in mmWave massive MIMO systems, where the number of antennas is very large and the transmission bandwidth is very wide. In this paper, a novel mmWave massive MIMO channel model is introduced which embraces the spatial-wideband effect. Then, a joint parameter learning algorithm is proposed to extract different channel parameters from the received data. Furthermore, a low-complexity spatial wideband channel estimation scheme is proposed for the new channel model. Simulation results and complexity analysis show that the proposed scheme are capable of achieving a considerable reduction in complexity compared with the recently spatial-wideband channel estimation algorithm with a small performance loss. Moreover, our scheme can effectively estimate the novel channel even part of the antenna data is received.