To program one or more non-volatile storage elements, a set of programming pulses are applied to at least one selected non-volatile storage element and one or more particular unselected non-volatile storage elements, for example, via a common word line. A boosting voltage is applied to other unselected non-volatile storage elements during the programming process in order to boost the channels of the unselected non-volatile storage elements so that programming will be inhibited. Each of the programming pulses has a first intermediate magnitude, a second intermediate magnitude and a third magnitude. In one embodiment, the first intermediate magnitude is similar to or the same as the boosting voltage. The second intermediate magnitude is greater than the first intermediate magnitude, but less then the third magnitude. Such an arrangement can reduce the effects of program disturb.