In South China during the hot season, which commonly lasts from June to October, weeds infest arable lands. The shifts in farming practices and climate change have brought new challenges to cropland weed management. Therefore, further research is needed to elucidate patterns of weed community occurrence during the hot season. To investigate weed community occurrence patterns and the effectiveness and subsequent influence of different weed control practices on weed occurrence, we conducted a series of field surveys and weed-control field experiments during the hot season on standard vegetable farms, as well as soil incubation studies in growth chambers under different temperature regimes. The experiments were conducted in Guangzhou city (located in South China) from June 10 to October 10, 2013. After a 50-day fallow period beginning on June 10, the weed community of the studied field was dominated by a few species of grass, with broadleaf weeds exhibiting high species diversity but lower coverage, and sedges having lower species diversity and coverage. Eleusine indica was mono-dominant among the 31 weed species observed in the experimental vegetable field. Echinochloa colona, Leptochloa chinensis, Cyperus iria, and Ludwigia hyssopifolia also showed high coverage. With respect to soil collected from the experimental field immediately after the fallow period, seedlings of grasses and sedges emerged and grew better under high temperature conditions, especially at 40 ℃/35 ℃, but were all strongly inhibited at 15 ℃. Seedlings of broadleaf weeds emerged and grew well at 15 ℃, while their growth was constrained at 40 ℃/35 ℃. After treatment with paraquat (900 g a.i./hm²), the trend of increasing dominance of broadleaf weeds in the experimental field was transformed into continuous mono-dominance of grasses. After application of paraquat followed by hoeing, weeds re-covered >60% of the soil surface after 30 days. In contrast, weed coverage was decreased by 72.9%, 84.6%, 83.9%, or 77.5% when followed by an application of the pre-emergence herbicides butachlor (1350 g a.i./hm²), acetochlor (750 g a.i./hm²), metolachlor (750 g a.i./hm²), or pendimethalin (600 g a.i./hm²), respectively. Furthermore, tillage to 15 cm after hoeing and before the application of pre-emergence herbicide greatly improved weed control effectiveness. Canonical correspondence analysis (CCA) suggested that the application of paraquat, hoeing, and tillage all showed significant (P < 0.01) influences on the weed community structure of the studied vegetable field 70 days after weed control treatments. After 70 days, paraquat application had strongly negative effects on the occurrence of Echinochloa colona and Leptochloa chinensis and strongly positive influences on the occurrence of Mimosa pudica, Corchorus aestuans, Ageratum conyzoides, and Portulaca oleracea. Hoeing and tillage had strongly negative influences on the occurrence of Eleusine indica, Echinochloa colona, Leptochloa chinensis, Digitaria sanguinalis, Fimbristylis miliacea, and Kyllinga triceps and strongly positive effects on the occurrence of broadleaf weeds. Common weed species in the experimental fields formed two clear groups in the CCA biplot with respect to the first two axes, namely (1) grass species plus two sedges, F. miliacea and K. triceps, and (2) broadleaf weed species plus Cyperus iria.