Nurse scheduling problem is a complex problem that often occurs in hospitals. Many constraints have to be fulfilled in order to maintain a fair schedule. Schedule maker must determine day-to-day shift arrangement of each nurse in a way that fulfills the constraints. Therefore, to solve this problem, research is conducted and software is made through this final project. The software is made using generic algorithm and heuristic search. A gene in a chromosome represents one nurse to be scheduled. A chromosom represents series of nurses to be scheduled. The initial population is generated randomly, then a heuristic search is performed. The next step is fitness evaluation. After calculating the fitness value, selection, crossover, and mutation are done sequentially using truncation selection, partially-matched crossover, and reciprocal exchange mutation method respectively. Then the heuristic search and fitness evaluation will be conducted again. The fittest chromosome at the end of the cycle will be selected as the solution and will be converted into the schedule. After series of testing are conducted, it can be summarized that the time needed to find the fittest chromosome is affected by the number of populations and number of generations. The chromosom with fitness value 1, meaning all nurses are successfully scheduled, can be obtained by using 10 populations and 10 generations, without considering the value of selection, crossover, and mutation probability. Fitness value is influenced by the sequences of gene in a chromosom.