Experiments on a transmission error, having a period of one revolution of the belt in polychloroprene rubber timing belts, were performed under a quasi-static condition and an initial tension. The experimental results were compared with the computed results obtained using the measurement results of the production error in the belt. It was confirmed that the transmission error was mainly caused by a single pitch error, a change in cross-sectional area of the load carrying cords and a change in modulus of elasticity for one revolution of the belt. As the experimental results coincided quite well with the computed ones, it was concluded that the measurement methods for the single pitch error and the belt elongation factor, and the arrangement methods of these values used to compute the transmission error were right. Further, it was found that the transmission error due to the production error in the belt lessens when the initial tension becomes smaller.