Atomic layer deposition (ALD) technology is a self-limiting ultrathin film deposition method. In ALD reactions, the reaction temperature plays a key role. And generalized predictive control (GPC) algorithm is adopted to control the temperature of a radiant heating reactor. The mathematical model of the system’s temperature uniformity is calculated with heat transfer theory. Since temperature control is affected by GPC parameters including soften factor, weighting coefficient, sample time etc, it is necessary to study the effect of these parameters. By experiments, the values of soften factor, weighting coefficient and sampling time could be optimized and then be applied in the process of ALD reactions. And conclusions can be draw that soften factor has the largest impact on the algorithm while the influences of weighting coefficient and sampling time are almost negligible. Efficient temperature control of the ALD reactor chamber using GPC algorithm with optimal parameters is demonstrated.