This paper proposes two networked iterative learning control(NILC)schemes for a class of linear-time-invariant single-input-single-output systems with stochastic communication delays being subject to Bernoulli-type distribution.In the two NILC schemes,the actual system input is the synchronous system input at the previous iteration if the corresponding system input at the current iteration is delayed,otherwise the actual system input is a linear combination of the synchronous system inputs at the current and previous iterations where the coefficients are dependent upon the input communication delay probability; for the output signals used by the ILC unit,we develop two strategies: one is the same as that for the actual system input,and the other one is that the actually utilized output is the synchronous desired output if the system output is delayed,otherwise the actually utilized output is a linear combination of the synchronous system output at current iteration and the synchronous desired output where the coefficients are dependent upon the output communication delay probability.It is shown that under certain conditions the system output is convergent to the desired output in the sense of expectation.Numerical experiments illustrate the effectiveness of the NILC schemes.