As a NP-hard combinatorial optimisation problem job-shop scheduling can be solved quite well by simulated annealing (SA) algorithm. tHis paper prsents the aplikation of schrage algorithm to subtitute random generation method of initial solution in SAalgorithm to increase its performance in terms of solution cuality and camputation time. One-machine schedules produced by schrage algorithm are combined to form an initial job-shopsc hedule. Program runs prove that this modification yields a better result in solving job-shop problem. The strong point of schrage algorithm is its nature to take into account the order of head and tail value of each operation to build a good initial ane-machine schedule.