In Australia, efforts are under way to revive the pigeonpea (Cajanus cajan (L.) Millsp.) industry, which has high export potential because of an increased demand in the international market. However, weeds are a major constraint to achieve high yield in pigeonpea. This study was designed to assess the effect of row arrangement and herbicide treatment on weed suppression and pigeonpea grain yield. Row arrangements included row spacing (narrow, 25 cm; wide, 50 cm) and paired rows (rows 25 cm apart within a pair, each pair separated from the next by 75 cm). Herbicide treatments were: untreated control; pre-emergent pendimethalin at 910 g a.i. ha^–1; post-emergent imazapic at 84 g a.i. ha^–1; and pre-emergent pendimethalin followed by post-emergent imazapic (rates as above). In the first year, Trianthema portulacastrum was the dominant weed, and infestation was 100% in the non-treated control plots. In the second year, other weeds (Setaria viridis, Eragrostis cilianensis and Chloris virgata) comprised 30% of the weed population. Averaged over row arrangements, grain yield varied from 2088 to 2689 kg ha^–1 in 2017 and from 835 to 2145 kg ha^–1 in 2018, and was lowest in the untreated control and highest in the plots treated with the sequential application of pendimethalin and imazapic. Averaged over years and herbicide treatments, yield was lower in paired rows (1850 kg ha^–1) than in narrow (2225 kg ha^–1) and wide (2165 kg ha^–1) row spacings. In the first year, all herbicide treatments provided >50% control of T. portulacastrum in the narrow and wide row spacings and increased yield by >22% over the untreated control. In the second year, the single application of imazapic proved inferior for controlling weeds, resulting in a 21% reduction in grain yield compared with sequential application of pendimethalin and imazapic. In both years, grain yield was similar for the single application of pendimethalin and sequential application of pendimethalin and imazapic. Despite the complex weed flora in 2018, the single application of imazapic provided acceptable weed control only when the crop was planted at 25 cm row spacing. Our results suggest that the single application of pendimethalin was effective on T. portulacastrum. However, in a complex weed flora situation, the sequential application of pendimethalin and imazapic provided effective weed control and resulted in improved yield.