This article implements the thinning process algorithm, which has been generalized for estimators of compound periodic Poisson processes. The use of generalizations in the algorithm has been prepared with a linear trend in the periodic elements. This research aims to discuss estimators of the variance function. The method used in this research is the simulation method. Simulation results using a generalized algorithm thinning process show that in the case of a limited observation time interval, some estimators are good enough to approach the actual value. As the value of n increases, the simulated value of the estimator moves towards the predicted value. This is following the lemmas, theorems, and consequences that have been discussed. It was also found that several estimators were quite slow. This results in the movement of the bias, variance, and MSE values of the estimators being slow, even though they are moving towards 0. So that further modifications can be made to the model being studied.