Background: The Extension of Community Healthcare Outcome program (ECHO) is an educational and training telemedicine service that provides voluntary case-based learning opportunities for healthcare professionals. Prediction of participant attendance will be useful to improve the sessions with early requirements to maximize the benefits. Usually, the most sophisticated Auto Regressive models are used for forecasting, when the data contains observation with multiple variables and dependencies, Simple Moving Average (SMA) models are used less in such conditions. In this study we want to examine the accuracy and reliability of Moving Average models compared with Auto Regressive models. The objective of this work is to develop an accurate forecasting model for ECHO program attendance by considering non-stationary, independent organizational data. Methods: The study analyzed 2015-2019 Show Me ECHO attendance data from the Missouri Telehealth Network (MTN). The first step; trained and tested both SMA and ARIMA predictive models without any dependent variables and evaluated both models by measuring error values. The second step; used the best model to forecast ECHO attendance for years 2020 - 2025. Results: The SMA model was better than the ARIMA model for independent data with lower error values MAE - 38.9, MSE - 2552.15, MAPE - 32.9 percent, p- value: 3.36E-28, and higher R - square: 87 percent. Where ARIMA model was with higher error values MAE - 61.8, MSE - 7198.88, MAPE - 37.7 percent, p- value: 6.25E-22, and lower R - square: 80 percent. Conclusion: Simple Moving Average (SMA) is more accurate than Autoregressive Integrated Moving Average (ARIMA) in forecasting future ECHO program attendance. Based on prediction; In 2019, the attendance range was 250-550, where in 2025 it got increased to 530-1170; shows that telehealth attendance will be doubled in th