Moral naturalism is the view that moral properties, such as lightness and goodness, are in some important sense "natural" properties. Some nat uralists have sought to make good on this idea by deploying a kind of semantics proposed by Hilary Putnam and others. Putnam himself sugges ted this strategy,1 and Richard Boyd has pursued it.2 The kind of semantics proposed by Putnam allows for true synthetic property identity statements, such as that water is H20. If this approach can be applied to moral pre dicates, it therefore opens the door to a kind of "synthetic semantic moral naturalism" according to which, for any moral property, there is a syn thetic truth to the effect that the property is identical to a certain natural property. This strategy has come under attack, however, in an argument by Terence Horgan and Mark Timmons known as the "Moral Twin Earth argument".3 On the one hand, Horgan and Timmons argue, G. E. Moore's "open question argument" defeated the idea that there are any analytically true sentences to the effect that a given moral property is identical to a natural property. On the other hand, they argue, the Moral Twin Earth argument undermines the idea that there are any synthetic truths to the effect that a given moral property is identical to a natural property. If they are correct, naturalism is in trouble.4 Are they correct? I will argue that they are not. A close look shows that the Moral Twin Earth argument poses no threat to moral naturalism. To say this is not to say that synthetic moral naturalism is without difficulties. I believe that in fact it has difficulty accounting for the normativity of moral claims. My goal in this paper, however, is the modest one of making Moral Twin Earth safe for naturalists.