A算法是一种启发式搜索算法，常用于路径规划等问题中。在使用A算法时，需要设置终止策略，以避免算法无限循环或者超出规定的运行时间。

一种常见的终止策略是基于节点的限制。即，给定一个最大节点数，当已生成节点数达到这个限制时，终止算法的运行。

下面是Python代码示例：

def A_star(graph, start, goal, max_nodes):
    open_list = [start]
    close_list = []
    g_score = {start: 0}
    f_score = {start: heuristic(start, goal)}

    while open_list:
        current = min(f_score, key=f_score.get)
        if current == goal:
            return path

        open_list.remove(current)
        close_list.append(current)

        if len(close_list) > max_nodes:
            return None

        for neighbor in graph[current]:
            if neighbor in close_list:
                continue

            tentative_g_score = g_score[current] + dist(current, neighbor)

            if neighbor not in open_list:
                open_list.append(neighbor)
                tentative_is_better = True
            elif tentative_g_score < g_score[neighbor]:
                tentative_is_better = True
            else:
                tentative_is_better = False

            if tentative_is_better:
                g_score[neighbor] = tentative_g_score
                f_score[neighbor] = g_score[neighbor] + heuristic(neighbor, goal)
                came_from[neighbor] = current

    return None

在上述代码中，max_nodes参数指定了节点数限制。如果已经生成的节点数超过这个限制，算法就会终止并返回None。这样，就可以避免算法无限循环或者超时的情况。