In this paper, we present a fixed-time convergent distributed cooperative learning algorithm for stochastic configuration network, namely FixD-SCN, to solve ’Big Data’ problem. The algorithm is proposed by using the fixed-time stability. In SCN, parameters that involve input weights and biases of new hidden nodes, are obtained at random via a supervisory mechanism with inequality constraints. And in particular, compared to the finite- time algorithms, the optimization of output weights are achieved within fixed time which is independent of the initial conditions of agents. Different from the traditional centralized learning algorithm, the FixD-SCN requires that each agent only cooperates with its nearest neighbors to share local information and learn unknown patterns. Numerical examples given corroborate effectiveness of the FixD-SCN algorithm.