diff --git a/internal/ui/practice_session.go b/internal/ui/practice_session.go
index 3701fc2..91115ef 100644
--- a/internal/ui/practice_session.go
+++ b/internal/ui/practice_session.go
@@ -341,7 +341,7 @@ func (m *PracticeSession) handleAnswerSubmission() (tea.Model, tea.Cmd) {
 	originalItem := m.getCurrentRawItem()
 	expectedInput := m.getExpectedInput(originalItem)
 
-	isCorrect := m.isInputCorrect(userInput, expectedInput)
+	isCorrect := m.isInputCorrectForItem(userInput, originalItem)
 	m.recordSpacedRepetition(originalItem, isCorrect)
 
 	if isCorrect {
@@ -972,6 +972,55 @@ func (m PracticeSession) getExpectedInput(item string) string {
 	return item
 }
 
+func (m PracticeSession) isInputCorrectForItem(userInput, item string) bool {
+	acceptedInputs := m.getAcceptedInputs(item)
+	for _, expected := range acceptedInputs {
+		if expected == "" {
+			continue
+		}
+		if m.isInputCorrect(userInput, expected) {
+			return true
+		}
+	}
+	return false
+}
+
+func (m PracticeSession) getAcceptedInputs(item string) []string {
+	expected := strings.TrimSpace(m.getExpectedInput(item))
+	trimmedItem := strings.TrimSpace(item)
+	_, translation := practice.ParseLine(item)
+	translation = strings.TrimSpace(translation)
+
+	acceptedSet := make(map[string]struct{})
+	if expected != "" {
+		acceptedSet[expected] = struct{}{}
+	}
+	if trimmedItem != "" {
+		acceptedSet[trimmedItem] = struct{}{}
+	}
+
+	if expected != "" && translation != "" {
+		withSpace := strings.TrimSpace(expected + " " + translation)
+		if withSpace != "" {
+			acceptedSet[withSpace] = struct{}{}
+		}
+		withArrow := strings.TrimSpace(expected + " ->> " + translation)
+		if withArrow != "" {
+			acceptedSet[withArrow] = struct{}{}
+		}
+	}
+
+	if len(acceptedSet) == 0 && trimmedItem != "" {
+		return []string{trimmedItem}
+	}
+
+	result := make([]string, 0, len(acceptedSet))
+	for candidate := range acceptedSet {
+		result = append(result, candidate)
+	}
+	return result
+}
+
 // 检查输入是否正确
 func (m PracticeSession) isInputCorrect(userInput, expectedInput string) bool {
 	// 获取匹配模式
diff --git a/internal/ui/practice_session_test.go b/internal/ui/practice_session_test.go
index 7574f9b..4c28541 100644
--- a/internal/ui/practice_session_test.go
+++ b/internal/ui/practice_session_test.go
@@ -192,6 +192,75 @@ func TestIsInputCorrect(t *testing.T) {
 	}
 }
 
+func TestIsInputCorrectForItem(t *testing.T) {
+	setupPracticeSessionTest(t)
+
+	session := &PracticeSession{}
+	item := "People often use the phrase '2' in conversation. ->> In everyday conversations, people frequently mention 'a great number of' when they want to emphasize quantity."
+
+	config.AppConfig.CorrectnessMatchMode = "exact_match"
+	exactMatchCases := []struct {
+		name  string
+		input string
+		want  bool
+	}{
+		{
+			name:  "只输入原文",
+			input: "People often use the phrase '2' in conversation.",
+			want:  true,
+		},
+		{
+			name:  "输入原文和翻译",
+			input: "People often use the phrase '2' in conversation. In everyday conversations, people frequently mention 'a great number of' when they want to emphasize quantity.",
+			want:  true,
+		},
+		{
+			name:  "输入带箭头的完整原始行",
+			input: "People often use the phrase '2' in conversation. ->> In everyday conversations, people frequently mention 'a great number of' when they want to emphasize quantity.",
+			want:  true,
+		},
+		{
+			name:  "仅输入翻译",
+			input: "In everyday conversations, people frequently mention 'a great number of' when they want to emphasize quantity.",
+			want:  false,
+		},
+	}
+
+	for _, tt := range exactMatchCases {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := session.isInputCorrectForItem(tt.input, item); got != tt.want {
+				t.Errorf("isInputCorrectForItem() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+
+	config.AppConfig.CorrectnessMatchMode = "word_match"
+	wordMatchCases := []struct {
+		name  string
+		input string
+		want  bool
+	}{
+		{
+			name:  "忽略标点仍判定正确",
+			input: "People often use the phrase 2 in conversation In everyday conversations people frequently mention a great number of when they want to emphasize quantity",
+			want:  true,
+		},
+		{
+			name:  "单词错误仍判定错误",
+			input: "People often use the phrase two in chat. In common talks, folks often mention a great number when they want to emphasize quantity.",
+			want:  false,
+		},
+	}
+
+	for _, tt := range wordMatchCases {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := session.isInputCorrectForItem(tt.input, item); got != tt.want {
+				t.Errorf("isInputCorrectForItem() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+}
+
 // 测试normalizeForWordMatch方法
 func TestNormalizeForWordMatch(t *testing.T) {
 	// 创建测试会话
@@ -245,9 +314,9 @@ func TestGetCurrentItem(t *testing.T) {
 
 	// 创建测试会话
 	session := &PracticeSession{
-		resourceType: practice.Words,
-		items:        []string{"apple ->> 苹果", "banana ->> 香蕉", "orange ->> 橙子"},
-		practiceOrder: []int{0, 1, 2},
+		resourceType:   practice.Words,
+		items:          []string{"apple ->> 苹果", "banana ->> 香蕉", "orange ->> 橙子"},
+		practiceOrder:  []int{0, 1, 2},
 		completedCount: 0,
 	}
 
@@ -357,4 +426,4 @@ func TestGetCurrentItem(t *testing.T) {
 	if result != "" {
 		t.Errorf("超出范围 getCurrentItem() = %v, want empty string", result)
 	}
-}
\ No newline at end of file
+}
diff --git a/mllt-cli b/mllt-cli
index 4398fce..ae3f930 100755
Binary files a/mllt-cli and b/mllt-cli differ