In many recent studies, the value of forest inventory information in the harvest scheduling has been examined. Usually only the profitability of measuring simultaneously all the stands in the area is examined. Yet, it may be more profitable to concentrate the measurement efforts to some subset of them. In this paper, the authors demonstrate that stochastic optimization can be used for defining the optimal measurement strategy simultaneously with the harvest decisions. The results show that without end-inventory constraints, it was most profitable to measure the stands that were just below the medium age. Measuring the oldest stands was not profitable at all. It turned out to be profitable to postpone the measurements until just before the potential harvests. Introducing a strict end-inventory constraint increased the number of stands that could be profitably measured. In this case, also the length of the planning horizon had a clear effect on what stands were profitable to measure. With a 15-year planning horizon, measuring the oldest stands was profitable while with longer planning horizons it was not. The interest rate did not affect the number of stands measured much, but it had a clear effect on the timing of the measurements.