Abstract: The assumption of homoscedasticity is commonly concerned in regression analysis. The assumption is not always appropriate in theory and application. In linear and nonlinear regression models, there have been many testing results to discuss homoscedasticity. Based on Wei（1995）, this paper deals with heteroscedasticity in nonlinear regression models with random weighted variance function. The likelihood ratio test and score test are obtained to test hypothesis of homoscedasticity. If heteroscedasticity exists, a estimating method of random weights is given.