The purpose of this study is to determine whether there is Contribution of work discipline, work motivation, and teacher perceptions about headmaster leadership style towards teacher performance at SMP N with National Standard School (SSN) Semarang. This study includes quantitative expo facto research. The population in this study were teachers at SMP N with National Standard School (SSN) Semarang counted 302. The samples in this study are 161 teachers obtained from the table that was developed by Isaac and Michael. The determination of sampling in this study uses with a sample random sampling. The data are collected by survey or questionnaire. Data analysis technique uses multiple linear regression analysis of test preconditions data that is normality test, multicollinearity, heteroscedasticity, autocorrelation test and test. The results showed: (1) there is a contribution to the discipline of work, work motivation and teacher perceptions about headmaster leadership style towards teacher performance at SMP N with National Standard School (SSN) Semarang, as shown by the acquisition of the F test with probability value 0.000 <0.05. From the results of data analysis, the R2 value counted 0.429 it means that there are contributions of work discipline, motivation, and teacher perceptions about headmaster leadership style towards teacher performance at SMP N with National Standard School (SSN) Semarang counted 42.9%; 2) there is a contribution of work discipline towards teacher performance at SMP N with National Standard School (SSN) Semarang, this is indicated with the acquisition probability value <0.05 with the effective contribution counted 12.82%; 3) there is a contribution of work motivation towards teacher performance at SMP N with National Standard School (SSN) Semarang, it is shown by the acquisition probability value <0.05 with the effective contribution counted 12.86%; 4) there is contribution of teacher perceptions about headmaster leadership style towards teacher performance at SMP N with National Standard School (SSN) Semarang, this is indicated with the acquisition probability value <0.05 with the effective contribution counted 17.25% and the classical assumption test results showed that the regression model is not biased or problems of classical assumptions (normality, linearity, and multicollinearity) that could otherwise BLUE (Best, Linear, Unbiased, Estimator).