Quantitative and qualitative changes in the weed community and the weed seed bank in the soil due to crop relation and weed control measurement were studied by the UNA Managua al ''Las Mercedes'' Experimental Station from 1987 to 1990. A bifactorial permanent field trial combined five grain crop relations and three weed management systems. To determine the weed seed bank in the soil a modified cultivation method was used. Results are showing significant diifferences between crop rotations and weed control systems after the first three year trial period. The level of weed infestation in the field (actual) varied from 88 to 440 ind./m(2), while the weed seed bank in the soil (potential) was determined from 3125 to 12959 secds/m(2), giving emergence rates from 0.5 to 14.2 %. The highest values were registered in the cucumber-soya rotation with 330 ind./m(2) and 6771 seeds/m(2). The chemical weed control with 217 ind./m(2) and 6063 secds/m(2) had a low efficiency due to predominance of tolerate species Rottboellia cochinchinensis, counting until 405 ind./m(2), The lowest actual infestation was reached in the corn-sorghum rotation with 101 ind./m(2) and the lowest potential in the corn-soybean relation with 4010 seeds/m(2), Periodic weed control reduced weed infestation until 109 ind./m(2) and 3531 seeds/m(2).