diff --git a/lnwallet/wallet.go b/lnwallet/wallet.go
index 5d2646bf..a003628f 100644
--- a/lnwallet/wallet.go
+++ b/lnwallet/wallet.go
@@ -1428,6 +1428,8 @@ func coinSelect(feeRate SatPerKWeight, amt btcutil.Amount,
 		//
 		// TODO: Handle wallets that generate non-witness change
 		// addresses.
+		// TODO(halseth): make coinSelect not estimate change output
+		// for dust change.
 		weightEstimate.AddP2WKHOutput()
 
 		// The difference between the selected amount and the amount
diff --git a/lnwallet/wallet_test.go b/lnwallet/wallet_test.go
new file mode 100644
index 00000000..2c2a21cc
--- /dev/null
+++ b/lnwallet/wallet_test.go
@@ -0,0 +1,180 @@
+package lnwallet
+
+import (
+	"testing"
+
+	"github.com/btcsuite/btcutil"
+	"github.com/lightningnetwork/lnd/input"
+)
+
+// fundingFee is a helper method that returns the fee estimate used for a tx
+// with the given number of inputs and the optional change output. This matches
+// the estimate done by the wallet.
+func fundingFee(feeRate SatPerKWeight, numInput int, change bool) btcutil.Amount {
+	var weightEstimate input.TxWeightEstimator
+
+	// All inputs.
+	for i := 0; i < numInput; i++ {
+		weightEstimate.AddP2WKHInput()
+	}
+
+	// The multisig funding output.
+	weightEstimate.AddP2WSHOutput()
+
+	// Optionally count a change output.
+	if change {
+		weightEstimate.AddP2WKHOutput()
+	}
+
+	totalWeight := int64(weightEstimate.Weight())
+	return feeRate.FeeForWeight(totalWeight)
+}
+
+// TestCoinSelect tests that we pick coins adding up to the expected amount
+// when creating a funding transaction, and that the calculated change is the
+// expected amount.
+//
+// NOTE: coinSelect will always attempt to add a change output, so we must
+// account for this in the tests.
+func TestCoinSelect(t *testing.T) {
+	t.Parallel()
+
+	const feeRate = SatPerKWeight(100)
+	const dust = btcutil.Amount(100)
+
+	type testCase struct {
+		name        string
+		outputValue btcutil.Amount
+		coins       []*Utxo
+
+		expectedInput  []btcutil.Amount
+		expectedChange btcutil.Amount
+		expectErr      bool
+	}
+
+	testCases := []testCase{
+		{
+			// We have 1.0 BTC available, and wants to send 0.5.
+			// This will obviously lead to a change output of
+			// almost 0.5 BTC.
+			name: "big change",
+			coins: []*Utxo{
+				{
+					AddressType: WitnessPubKey,
+					Value:       1 * btcutil.SatoshiPerBitcoin,
+				},
+			},
+			outputValue: 0.5 * btcutil.SatoshiPerBitcoin,
+
+			// The one and only input will be selected.
+			expectedInput: []btcutil.Amount{
+				1 * btcutil.SatoshiPerBitcoin,
+			},
+			// Change will be what's left minus the fee.
+			expectedChange: 0.5*btcutil.SatoshiPerBitcoin - fundingFee(feeRate, 1, true),
+		},
+		{
+			// We have 1 BTC available, and we want to send 1 BTC.
+			// This should lead to an error, as we don't have
+			// enough funds to pay the fee.
+			name: "nothing left for fees",
+			coins: []*Utxo{
+				{
+					AddressType: WitnessPubKey,
+					Value:       1 * btcutil.SatoshiPerBitcoin,
+				},
+			},
+			outputValue: 1 * btcutil.SatoshiPerBitcoin,
+			expectErr:   true,
+		},
+		{
+			// We have a 1 BTC input, and want to create an output
+			// as big as possible, such that the remaining change
+			// will be dust.
+			name: "dust change",
+			coins: []*Utxo{
+				{
+					AddressType: WitnessPubKey,
+					Value:       1 * btcutil.SatoshiPerBitcoin,
+				},
+			},
+			// We tune the output value by subtracting the expected
+			// fee and a small dust amount.
+			outputValue: 1*btcutil.SatoshiPerBitcoin - fundingFee(feeRate, 1, true) - dust,
+
+			expectedInput: []btcutil.Amount{
+				1 * btcutil.SatoshiPerBitcoin,
+			},
+
+			// Change will the dust.
+			expectedChange: dust,
+		},
+		{
+			// We have a 1 BTC input, and want to create an output
+			// as big as possible, such that there is nothing left
+			// for change.
+			name: "no change",
+			coins: []*Utxo{
+				{
+					AddressType: WitnessPubKey,
+					Value:       1 * btcutil.SatoshiPerBitcoin,
+				},
+			},
+			// We tune the output value to be the maximum amount
+			// possible, leaving just enough for fees.
+			outputValue: 1*btcutil.SatoshiPerBitcoin - fundingFee(feeRate, 1, true),
+
+			expectedInput: []btcutil.Amount{
+				1 * btcutil.SatoshiPerBitcoin,
+			},
+			// We have just enough left to pay the fee, so there is
+			// nothing left for change.
+			// TODO(halseth): currently coinselect estimates fees
+			// assuming a change output.
+			expectedChange: 0,
+		},
+	}
+
+	for _, test := range testCases {
+		test := test
+		t.Run(test.name, func(t *testing.T) {
+			t.Parallel()
+
+			selected, changeAmt, err := coinSelect(
+				feeRate, test.outputValue, test.coins,
+			)
+			if !test.expectErr && err != nil {
+				t.Fatalf(err.Error())
+			}
+
+			if test.expectErr && err == nil {
+				t.Fatalf("expected error")
+			}
+
+			// If we got an expected error, there is nothing more to test.
+			if test.expectErr {
+				return
+			}
+
+			// Check that the selected inputs match what we expect.
+			if len(selected) != len(test.expectedInput) {
+				t.Fatalf("expected %v inputs, got %v",
+					len(test.expectedInput), len(selected))
+			}
+
+			for i, coin := range selected {
+				if coin.Value != test.expectedInput[i] {
+					t.Fatalf("expected input %v to have value %v, "+
+						"had %v", i, test.expectedInput[i],
+						coin.Value)
+				}
+			}
+
+			// Assert we got the expected change amount.
+			if changeAmt != test.expectedChange {
+				t.Fatalf("expected %v change amt, got %v",
+					test.expectedChange, changeAmt)
+			}
+		})
+	}
+}