To assess errors of a grab sample methods for measurement of radon-222 decay products concentration in air, a computer program simulating deposition of radon daughters on a filter and of variation of their activity against time was developed. Employing this program a measuring procedure of three selected methods of measurement were simulated: measurement of decay product alpha activity in three time intervals, in two time intervals and in one time interval. The count numbers registered in selected time intervals for RaA:RaB:RaC ratio from 200:200:200 ... 200:0:0 dis/min at counting efficiency h = 0.3 were processed according to the equations corresponding to the method under investigations and relative errors resulting from statistics of counting and from deficiency of the method were computed. The results of computations of two interval method were compared with practical measurements from international intercomparison of such gauges. Computations show that the three interval method eliminates errors of the method but large statistic error remains resulting from the uncertainty of measurement of alpha radiation in selected intervals reaching 9.4% relative of alpha potential energy. Two interval method exhibits errors resulting from deficiency of the measuring method 18.3% and from the statistics of pulse counting 7.3%. Measurements at international intercomparison of such gauges showed precision of measurement (statistical error plus method error) 12.6% of alpha potential energy. Computer simulation of one interval method showed 13.5% error related to the deficiency of the method and 3.4% statistic error.