A field experiment was conducted to investigate the direct, residual and cumulative effect of boron on wheat and rice yield in B deficient soil (0.29 mg kg-1) for two consecutive years. Three rates of B, ranging from 0 to 2 kg ha-1 in soil, were applied as borax (B: 11%) along with basal dose of nitrogen, phosphorus and potassium. The results showed that the direct application of 2 kg B ha-1 in wheat was more beneficial as compared to cumulative or residual application in rice-wheat system. Highest grain yield of wheat (4803 kg ha-1) was recorded with the application of 2 kg B ha-1 only to wheat crop (direct) while 4770 kg ha-1 recorded with the application of 2 kg B ha-1 both to wheat and rice (cumulative). The yield increased due to residual effect of B was statistically lower than the cumulative effect of B. In rice maximum paddy yield was recorded with the cumulative application of B followed by direct and residual applied B in rice-wheat system. Boron concentration in soils ranged from 0.1 to 0.58 mg kg-1 in wheat and 0.10 to 0.69 mg kg-1 in rice, while in leaves it ranged from 6-9.6 mg kg-1 in wheat and 8-16 mg kg-1 in rice. The concentration of B in soil and leaves increased due to the treatments in the order; cumulative > direct effect > residual > control (without B). The yield attributes like 1000-grain weight, number of spikes, spike length and plant height were increased by the residual, direct and cumulative effect of B levels. Under B-deficient soil conditions, yield of wheat can be increased by direct application of B, while yield of rice can be maximized by the cumulative use of B fertilization.