This study assessed varied types of Web service-based learning environments in a digital-filter design course. Ninety-one undergraduate students participated in an experimental study. Two independent variables were studied: (1) Web service-based simulations (batch-based, comparison-based, and interval change-based simulations); and (2) prior Internet familiarity (high and low). Two dependent variables were measured: (1) a knowledge achievement test, and (2) a problem-solving belief test. The experimental research design of the study was a 3 × 2 randomized post-test design. Multivariate Analysis of Variance (MANOVA) was used to analyze collected data. The main effects and the potential interaction of the two independent variables were examined. Results indicate that, if not considering extraneous variables, Web service-based simulation with a simple batch-based design yielded a significantly better learning performance than the other two complex simulation designs (comparison-based and interval change-based simulations) (F[2,85] = 4.274, p<.05).