This paper introduces a modified sampled-data control for solving the stabilized problem of stochastic complex multi-weights networks with time delay and Levy noise, where the sampled states depend on a time varying delay. That is, we design a class of delay sampled-data control to stabilize the given stochastic complex multi-weights networks with time delay and Levy noise in the sense of exponential stability in mean square. Using the Lyapunov method, graph theory, and some techniques of inequality, several types stabilized criteria are obtained. The control law is determined simultaneously, which depends on the perturbed intensity of noise, the coupling strength, time delays, and the lower and upper bounds of sampling intervals. In particular, as a practical application of our theoretical results, the stabilized problem of stochastic delayed oscillator networks with Levy noise and multi-weights is researched, and the correlative numerical simulations are provided for illustration.