This research aims to analyze how big the influence of CAR, NPF, FDR and BOPO against ROA Islamic banking in Indonesia. Approach to research with quantitative methods using secondary data. Types of quantitative data in the form of data runtun time (time series) monthly from the SPS-BI/OJK 2006:1-2018:1 period. Research models elaborated in the equation a linear multiple regression analysis with 0.05 significance level. From the results of the test, the variable F is free (CAR, NPF, FDR and BOPO) simultaneously significantly affect a variable is bound (ROA). This shows the regression model on CAR, NPF, FDR and BOPO can be used to predict the ROA. Whose hypotheses there is significant influence among variables a variable is bound against free. From the results of the test t, variable CAR influential and insignificant against the variable variable ROA, NPF has no effect and significantly to the variable variable FDR ROA, influential and significant variable against ROA. Then, the variable BOPO has no effect and does not significantly to variable ROA. From the results of the estimation model, the retrieved value R of the determination coefficient R 0,781 and the Square of the mean 0,610 of 61,10% of the variation could be explained by the bound variables are variables in the model. While remaining 38,90% explained by other causes which are not entered into the model.