The communication quality of the optical fiber network by cloud data scheduling greatly influence the balance, in order to improve the quality of optical fiber communication network, improve network cloud data transmission throughput and fidelity rate scheduling in parallel by the cloud data, cloud data transmission is proposed based on adaptive equalization, decision feedback equalization of optical fiber network cloud data parallel scheduling model construction. The transmission channel model of optical fiber network communication, quantitative estimation of optical fiber network by using the minimum mean square error estimation fusion, was treated by interference filter cloud data matched filter detector, the adaptive decision feedback equalization method for channel equalization, channel equalization in cloud data to the filter output of the parallel scheduling and multi thread output modulation. The simulation results show that using the method of parallel scheduling in cloud data in the fiber network is balanced well, the fidelity of the output data is high, and the bit error rate is low, which improves the overall performance of the optical network.