Nowadays,　navigation　systems　are　widely　used　in　public　travel　because　they　can　instantly　offer　GPS-based　route　directions.　Following　the　navigation　prompt　messages　while　driving　is　considered　a　secondary　driving　task,　while　vehicle　control　is　regarded　as　a　primary　driving　task.　Navigation　prompt　messages　with　more　information　can　deliver　more　cues　to　drivers,　but　they　require　a　higher　cognitive　demand　and　vice　versa.　To　systematically　explore　the　effects　of　the　amount　of　information　from　navigation　voice　prompts　and　further　quantify　the　utility　of　voice　prompts,　four　types　of　prompt　messages　with　increasing　amounts　of　information,　denoted　as　a　Single　Message,　Double　Message,　Triple　Message,　and　Quadruple　Message,　were　designed.　A　driving　simulation　experiment　was　conducted　to　obtain　driving　behavior　data　under　different　prompt　messages.　The　one-way　analysis　of　variance　(ANOVA)　and　Kruskal-Wallis　(KW)　test　were　used　to　examine　the　differences　in　driving　performance　under　the　guidance　of　different　prompt　messages　from　multiple　perspectives.　Then,　eight　indicators　were　selected　based　on　the　functions　of　the　navigation　system　and　the　driver＇s　response,　and　the　grey　near-optimal　method　was　used　to　determine　the　utility　of　the　four　types　of　prompt　messages.　This　study　found　that　the　four　types　of　navigation　prompt　messages　all　began　to　take　effect　at　about　200　m　upstream　of　the　stop　bar.　The　differences　between　the　four　types　of　prompt　messages　were　more　significant　in　the　zone　from　100　m　upstream　and　ended　at　100　m　downstream　of　the　stop　bar　of　the　intersection.　Drivers　using　Single　and　Double　Messages　exhibited　more　powerful　deceleration　than　those　using　Triple　and　Quadruple　Messages.　The　utility　values　of　the　four　types　of　prompt　messages　increased　with　the　increase　in　the　amount　of　information.　This　study　provides　theoretical　support　for　optimizing　navigation　information　and　lays　a　foundation　for　establishing　navigation　broadcast　guidelines.